# ICP: Mika

**Segment:** US-headquartered marketing/creative/digital agencies, 10–50 employees on LinkedIn, with a founder/owner/partner still publicly listed as principal, whose published client roster or case-study pages are majority B2B software/tech.

*Filter buildability note:* headcount, HQ, and founder-principal status are directly filterable (LinkedIn, Clutch, Semrush and HubSpot partner directories). **Majority-B2B-software roster is not available from any dataset**: the dossier states plainly that no source segments agencies by client vertical, so this attribute must be hand-verified from portfolio pages. *"Sells strategy alongside execution retainers"* is deliberately **not** a membership filter. It is a need condition in firmographic costume, and it excludes by definition any agency that wouldn't want the product. Treat it as the first qualifying question on the call, not a list attribute.

**Segment brief:** A **boutique B2B SaaS agency** delivers execution across some mix of content, paid, SEO, social, PR, or design, and is increasingly asked for **the strategic layer** on top: positioning, GTM sequencing, messaging frameworks. The structural pressure is economic, not creative: clients frequently "opt to skip strategic planning in favor of immediate implementation," while the internal cost of doing it properly is, in one agency principal's words, "tens of thousands of dollars in internal resources and tons of time," and the agency's own comp structure "does not reward non-billable activities." So the work is real, senior, and unfunded as a line item. These shops already live in Notion, Google Docs, Ahrefs, HubSpot, Apollo, Asana, ChatGPT and Claude; one documented agency has built a custom GPT per client to approximate what this product sells (n=1 in the dossier, Develomark; do not present as a segment-wide pattern). They will buy because the constraint is not writing capacity but access to a senior second opinion that argues back before a client does, at a price that works for a routine account rather than the flagship one: the function a fractional CMO or positioning consultancy performs (marketplace-quoted at $150–500/hr and $4k–20k/mo retainers; T2 vendor pages, no stated n). They will not buy on speed or volume claims, because the thing they fear most is producing work that sounds like everyone else's.

**Persona:** The agency founder. Founder, Owner, or Managing Partner of a 10–50 person US B2B SaaS marketing agency, still personally on new-business pitches and strategy calls.

**Persona brief:** They buy because in a founder-led shop the person doing the strategic thinking and the person signing a small-tool invoice are the same person, which is why the brief describes a founder-led sale with no procurement. They are accountable for new business won, retainers renewed, and margin. At 13% average after-tax margin, senior unbillable hours are the expense they feel most directly. Personal failure is specific and commercial: standing in a pitch unable to explain why their agency is different from a cheaper alternative, or receiving client feedback that is "clearly from ChatGPT," an inversion the dossier flags as real and first-person. They are not AI-skeptical; they use ChatGPT daily and can spot its output "from a mile away," which is precisely the objection. What they need and do not have is a senior peer who will pressure-test a positioning angle at speed and tell them they are wrong, at a latency and price that works for a routine account. They will pay for judgment and challenge; they will not pay for another content generator, and they will churn silently if the tool becomes a second system to maintain. *(inference: the "cannot delegate strategy" framing is deliberately absent. The dossier's one live agency job posting advertises "our AMs lead strategy," so delegation demonstrably happens in shops this size.)*

**Core tension:** **Strategy is the most senior work in the shop and the least billable, so it gets absorbed into unfunded founder hours or skipped, and the obvious relief valve, AI, is the one thing this founder and their clients can already recognise the output of.**


# How This Profile Was Built

This profile was assembled from live public research conducted 2026-08-18 and the client's brief. Sources include Myosin's own product pages and the published OpenClaw skill spec, competitor positioning copy, adjacent-category review text on G2 and Capterra, practitioner discussion in Glassdoor community bowls and Sharebird, agency and PMM job postings, and published firmographic reports (Promethean Research's 2026 State of Digital Services, n=119). No buyer interviews were conducted, and no Hivemind customer, case study, review, or quantified outcome exists at any source tier. Items tagged `[E]` trace to found evidence with a locatable source. Items tagged `[I]` are analyst inference built on that evidence: they are structurally reasoned, not observed, and should be treated as hypotheses to pressure-test in the first ten sales conversations rather than as findings to budget against.

**Validate first:**

1. **That the buyer is a US boutique B2B SaaS agency founder at all.** Every public Hivemind surface is web3-scoped, including the documented corpus ("a curated Web3 marketing knowledge base") and both parent-agency case studies. The agency ICP appears in the client's brief and in one FAQ disclaimer. If the first ten conversations skew crypto, the segment is a pivot target, not a customer base, and every downstream section forks.
2. **That founders will pay for pushback rather than output.** The adversarial mechanic is the only claimed differentiator that survives the parity list, and it rests entirely on one vendor-authored demo exchange. Test whether a founder describes wanting to be argued with, or quietly wants a faster first draft. Those are different products at different prices.
3. **That the unfunded strategic layer lands on the founder specifically.** Promethean's 68% is survey-sample composition, not a population estimate, and a live agency job posting advertises "our AMs lead strategy." If senior ICs absorb this work, the buyer, the budget line, and the sales cycle all move.
# Ideal Customer Profile

### Company Type

- **B2B (services): The execution-retainer shop that got promoted into strategy:** 10–50 FTE, US, founder still on every new-business call, delivering content, SEO, paid, or design on monthly retainers for B2B software clients, and now fielding requests for the strategic layer on top. The seller's own agency-facing copy names the pattern: clients "keep asking for more: positioning, GTM synthesis, campaign architecture" `[E]`. The pain repeats monthly because the work is real and senior while the comp structure "does not reward non-billable activities" `[E]`, so the positioning thinking lands in the founder's evenings and never appears on an invoice. Readiness is observable from outside: a Clutch or Semrush partner profile listing execution services only, a case-study page whose logos are majority SaaS, and a founder posting about positioning or messaging on LinkedIn. They buy Hivemind as a second opinion that pushes back at the moment an angle needs pressure-testing, not as another content generator `[I]`. **Fit: 9/10.** This archetype is described in the seller's own verbatim copy and matches the dossier's stated best-customer segment `[E]`.

- **B2B (services): The positioning-led boutique that sells the strategy sprint as the front door:** smaller, often 10–25 FTE, founder is the named strategist, and the commercial motion is a fixed-fee positioning or messaging engagement that opens the door to a retainer. Their exposure to the slop risk is the highest of any archetype: the buyer can already spot AI copy "from a mile away" `[E]`, and this shop's entire premise is that its judgment is not interchangeable. **(Hidden gem)** Their fixed-fee sprint is the worst-margin product they sell, because the price is capped at signature while the senior hours it consumes are not. That's why the most differentiated offering in the shop is the one quietly subsidized by the retainers `[I]`. That makes them ready for a tool used privately, upstream of the client, where this audience already sanctions AI "mostly for strategy or operations, never for creative" `[E]`. Readiness signal: a productized sprint with a published price on the services page. **Fit: 8/10.** Strong inferential logic, but no dossier evidence separates this archetype's conversion rate from the first `[I]`.

- **B2B (services): The multi-account growth agency running fifteen-plus small SaaS brands at once:** account-manager-led, high client count, low revenue per account, founder acting as floating strategist across a book too wide to hold in one head. The pain is acute and the buying trigger is real. The constraint is the product, not the buyer: Hivemind ships no per-client memory, no client workspace, and no white-label or client-facing output mode `[E]`, so every conversation starts from cold context and the founder re-explains the account each time. This archetype's documented workaround is already sharper than the product for their shape of problem, with one agency building a custom GPT per client `[E]`. Sell them the challenge mechanic on one flagship account rather than portfolio coverage, and expect silent churn if they try to run the whole book through it `[I]`. **Fit: 5/10.** Capped by an inferred capability mismatch against explicitly documented gaps `[I]`.

### Target Industries

**Overarching industry (single best fit): Marketing, advertising, and creative services, specifically independent, founder-led agencies serving B2B software and tech clients.** Every retainer in this industry bills execution hours while the positioning, GTM sequencing, and messaging framework work rides along unpriced: clients "opt to skip strategic planning in favor of immediate implementation," and the agency's own comp model "does not reward non-billable activities." [E] The unfunded strategic layer isn't a quarterly event here. It recurs on every new logo, every renewal, and every pitch the founder personally attends. [I]

**1) AI-forward independent agencies already running custom GPTs, Claude Code, or MCP servers per client (10–40 FTE, B2B software roster).** Expected win rate 7/10, inference-capped, but adoption friction is structurally lowest here and the client work is squarely B2B software. **(Hidden gem)** Hivemind ships zero workflow integrations, no Slack, no Notion, no HubSpot, which reads as a fatal gap until you notice the one surface it *does* ship is an outbound agent skill that lets other agents query the corpus. [E] For an agency that already orchestrates work inside Claude Code or OpenClaw, the integration gap inverts: they don't need Hivemind to connect to their stack, because their stack connects to Hivemind. Every other archetype pays the missing-integrations tax; this one is the only slice that gets the API-first architecture as a feature rather than an omission. [I] Caveat it in the room: keys are "issued at the discretion of Myosin," the public skills repo has 1 star and 0 forks, so you are selling an early surface to early people. [E] `readiness signals include` a public GitHub org with prompt, CLAUDE.md, or MCP-server repos; blog or LinkedIn content describing a custom GPT built per client; job postings for "AI ops," "AI engineer," or prompt-engineering skills at an agency-classified employer. [E]

**2) B2B SaaS content and SEO retainer shops (10–50 FTE, Notion-run editorial operations).** Expected win rate 6/10, large, enumerable, reachable, and the densest concentration of majority-software rosters in the segment; docked because the closest competitor targets them directly. The felt pain is being paid for volume while the client keeps asking why the content isn't moving anything, which is a positioning question the retainer doesn't fund. [I] Their workaround is documented: "Notion is the center of our content operation… Every article starts in Notion," a system that reflects the founder's own assumptions back at them and never disagrees. [E] The wedge is the pressure-test before the brief, not another draft. Expect the parity wall: M1-Project already runs the ChatGPT-comparison play at this exact buyer, in customer voice. [E] `readiness signals include` a published client roster of majority SaaS/software logos on the agency site; Ahrefs or Semrush agency-partner directory listings; job postings naming Notion, Surfer, and Ahrefs in the same requisition. [E]

**3) B2B demand-gen and paid-media agencies serving PLG and mid-market SaaS (15–50 FTE).** Expected win rate 5/10, real need, but the buyer's instinct is to solve it with a hire or a fractional CMO, not a subscription. Media buying is the most price-comparable service in the industry, and the founder's exposure is the pitch moment where they cannot articulate why they cost more than a cheaper shop. [E] Their category anxiety is on record from their own peers: "you can't win a price war with someone whose costs are near zero." [E] Strategy is how they escape the comparison, and it is the thing they have least time to do. One live agency posting names the wound precisely: "Tired of managing 17 accounts and getting no strategic seat at the table?" [E] `readiness signals include` Google Partner or Meta Business Partner badges alongside B2B/SaaS case studies; Built In or LinkedIn postings listing Performance Max, LinkedIn Ads, and HubSpot together; recent hires with "strategy" or "account strategy" titles in a shop previously staffed only with buyers. [E]

**4) Brand and positioning studios serving B2B tech (5–25 FTE, sell strategy as the product).** Expected win rate 4/10, highest need, worst identity fit, and worth working only as a founder-to-founder conversation. Positioning *is* their billable deliverable, so a tool that generates it threatens the line item, and the slop risk is existential rather than cosmetic: this is the buyer who can spot AI copy "from a mile away" and hates it. [E] Compounding it, Hivemind's own price anchor argues against "the agency markup," so you are selling replace-the-strategist copy to people who are the strategist. [E] The only survivable frame is the audience's own permission structure: AI "mostly for strategy or operations, never for creative," sold as the second opinion that stress-tests the founder's angle before the client hears it, never as the thing that produces it. [E] `readiness signals include` service pages naming positioning sprints, messaging architecture, or brand strategy as a standalone paid engagement; founder appearances on 2Bobs or Build a Better Agency; Fletch PMM or April Dunford frameworks cited in the agency's own published content. [E]

**5) Web3, crypto, and frontier-tech marketing agencies (10–50 FTE, protocol and ecosystem clients, US-headquartered), PIVOT HYPOTHESIS, not a segment member.** Expected win rate 4/10 as a *test*, not a forecast: these agencies fall outside the segment definition (majority-B2B-software roster), and the crypto skew in Hivemind's public evidence is a signal the company's current base sits here, not that this is where the reported ICP should go. Rank it last and run it as a two-account experiment, not a territory. The reason it's on the list at all is a corpus bound the other four archetypes inherit: the documented knowledge base is "a curated Web3 marketing knowledge base" with trigger conditions on token launch tactics and community growth, and cross-industry applicability is only an FAQ assertion. [E] For archetypes 1 to 4 that is a caveat to carry into every demo; for this one it is a subject-matter match, which makes it the cleanest place to test whether the retrieval layer actually improves answer quality versus a general model. [I] `readiness signals include` Clutch or agency-site portfolios listing protocol, L2, DeFi, or DAO clients; agency-handle activity on Farcaster or Crypto Twitter; LinkedIn attendance posts for KBW, DevConnect, or ETHDenver. [E]
### Annual Revenue

- **Primary band: $1.5M–$7M total agency revenue (services revenue, not ARR: this segment bills retainers and projects, not subscriptions).** Derived from the only hard firmographic in the dossier: Promethean Research's 2026 agency survey (n=119, 74% US) puts the average respondent at 31 employees and $4.43M revenue, implying roughly $143k revenue per FTE, which maps a 10–50 headcount shop into this band `[E for the survey figures, I for the per-FTE derivation (Promethean's own revenue-per-FTE benchmark is paywalled and was NOT FOUND)]`. The band works from both ends. At the low end, the unfunded strategic layer is already expensive: at 13% average after-tax margin, a founder absorbing senior unbillable hours on positioning and pitch strategy is spending the scarcest margin in the business, and the market rate for the alternative is quoted at $150–500/hr for fractional CMOs and $4k–20k/mo retainers `[E, T2 marketplace pages with no stated n]`. At the high end, $7M is roughly where a shop stops being one calendar's worth of strategic capacity and starts having a strategy lead who owns the artifact. Inside the band the buying path is one person: 68% of Promethean's respondents were founders, owners or partners, and the brief describes a founder-led sale closing in 2–4 weeks with no procurement `[E]`. Hivemind publishes no price at any tier, so the "small-tool budget" fit is an assumption, not a finding. Comparable AI strategy tools sit at $45–$99/mo, and anything above roughly $500/mo pushes the founder into comparing against a human strategist instead of against ChatGPT `[E for competitor prices; I for the threshold]`.

- **Adjacent band to test selectively: under $1.5M (sub-10 FTE shops).** This is not a fringe test. Promethean puts 88% of the 50,000+ US and Canadian agencies under 10 FTE, meaning the primary band is the top ~12% of the population and the volume sits below it `[E]`. What changes: the founder is not just the final strategist but the only one, so the "second opinion that pushes back" pitch gets *stronger*, while price sensitivity gets sharper and the free-ChatGPT default gets harder to dislodge. Same channels reach them (Grow Your Agency Slack is $35 lifetime, Sakas office hours are free), so the test is cheap to run `[E]`. (Hidden gem) The tell that separates a buyer from a tire-kicker here is not revenue but whether the shop has ever *charged separately* for a positioning or GTM engagement. An agency that has invoiced strategy as a line item has already decided the work is worth money and is buying capacity, an agency that has only ever given it away inside a retainer is buying a way to keep giving it away faster, and will churn when the free alternative improves `[I]`.

- **Avoid: $20M+ / 100+ FTE agencies and holdco-affiliated shops.** Three specific blockers. Procurement weight arrives with data governance questions Hivemind cannot answer today: no SOC 2, no DPA, no training-opt-out statement exists on any surface, and Digiday reports that ownership, access rights and portability have moved from procurement into standing governance conversations at this tier `[E]`. The buyer changes: the signer is a chief strategy officer or head of AI, not a founder on a pitch, and the evaluation becomes a platform comparison Hivemind loses on integrations and export `[I]`. And it is the competitor stronghold by construction: this is where Omni, WPP Open and Marcel already sit, and where "agentic / end-to-end / AI-powered" language is documented as burned to the point of meaninglessness `[E]`.


### Company Size

- **15–40 employees (primary).** Function shape: a founder plus one to three people who can credibly run a positioning conversation, usually senior account directors or an ops-heavy partner, with zero or one dedicated strategy title; the other 12–35 are execution: content, paid, SEO, design, AM. That shape produces the unfunded strategic layer directly: at this size the shop carries roughly a dozen or more retainers, so strategic asks arrive weekly, while Promethean's 2026 survey (n=119, avg 31 FTE, avg $4.43M revenue, 74% US) puts average after-tax margin at 13%, down from a ~15% long-run norm `[E]`. There is no slack to fund a senior head whose output nobody bills, so the residual lands on the founder's non-billable hours `[I]`. Identified from outside: LinkedIn "11–50 employees" plus US HQ plus a founder, owner, or managing partner still listed as principal; cross-check on Clutch or the Semrush and HubSpot partner directories; then hand-verify the B2B software roster from case-study pages, because no dataset segments agencies by client vertical `[E]`. The team-page tell is a headcount in the twenties or thirties with at most one "Strategy" title on it `[I]`. **(Hidden gem)** The account director is the buying committee member nobody names. One live agency posting in this band recruits on the line "our AMs lead strategy," which means strategy is already delegated at 15–40 heads `[E]`. The founder's actual problem is not capacity, it is quality control over strategy they did not write, and the tool gets bought as a way to make a mid-level strategic answer survive a founder's read before it reaches the client `[I]`. Sell to the founder, but ask on the call who currently drafts the positioning.

- **8–15 employees (secondary).** Function shape: the founder is the entire strategic bench, no strategy title exists, and the next-most-senior person is a doer with client-facing polish. The pain is more acute and the budget is thinner. Every hour of positioning work is founder time competing against pitching, hiring, and payroll, so the failure mode here is skipping the strategic layer entirely rather than absorbing it `[I]`. Volume is enormous and quality is uneven: Promethean counts more than 50,000 agencies across the US and Canada with 88% under 10 FTE, so this band shades quickly into freelancer collectives that will never pay for a second opinion `[E]`. Identified from outside: LinkedIn's "2–10" and lower "11–50" buckets are too coarse to separate a real 12-person shop from a three-person shop with contractors, so count named humans on the team page and cross-reference employee-listed profiles rather than trusting the filter `[I]`. Qualify on retainer language and named B2B software clients, not on headcount alone.

- **40–75 employees (upper bound, selective).** Qualifying condition: the founder or managing partner is still personally on new-business pitches and there is no Chief Strategy Officer, Head of Strategy, or Director of Operations on the team page. Above roughly 40 heads, strategy usually gets institutionalized into a department with its own leader, and that leader is the wrong buyer twice over: professionally invested in defending the function, and senior enough to trigger a real evaluation `[I]`. Hivemind cannot survive that evaluation today: no published pricing at any tier, no SOC 2 or DPA, no training-opt-out statement, and no case study, named customer, or review-site presence at any source tier `[E]`. The band stays reachable only while the signer is still the strategist, which is also what preserves the fast, uncommitteed sale the brief describes `[I]`. Identified from outside: LinkedIn "51–200" filtered to US HQ, then disqualified on the presence of any strategy-department or operations-leadership title, and on job-post volume that shows recruiting for a strategy function rather than for delivery `[I]`.
### Geographic Location

**United States, treated as a hard boundary rather than a starting point.** Promethean Research's 2026 agency survey (n=119) puts 74% of respondents in the US, with an average of 31 employees and $4.43M revenue: the exact center of the 10–50 FTE band [E]. One caveat governs everything below: no dataset segments agencies by client vertical, so metro ranking here is a proxy built from marketing talent pools, B2B SaaS client clusters, and the segment's own event calendar, not a count of agencies with majority-B2B-software rosters [E]. Descending order of density: **New York**, the deepest marketing talent pool in the country at roughly 410K marketing professionals per LinkedIn data cited by the AMA, and the home of Mirren Live, the agency new-business conference [E]. **San Francisco Bay Area**, where investor density puts the Series A–B B2B SaaS client base within driving distance of the agency serving it, and where PMA's Launch runs January 21, 2027 [E]. **Los Angeles**, second-largest marketing labor market at ~190K, though its roster skew is consumer and entertainment, so hand-verification of portfolio pages matters more here than anywhere [E/I]. **Boston**, a genuine B2B software cluster plus MarketingProfs B2B Forum, November 2–4, 2026 [E]. **Chicago**, where Built In Chicago's agency-filtered board is deep enough to read installed stacks straight off live postings [E]. **Dallas**, an AMA top-five marketing metro [E]. **Denver**, which is functionally the segment's event capital: AgencyCon October 8–9, 2026, PMA's Product Marketing Summit April 7–8, 2027, and AMI's Build a Better Agency Summit at the Westin Denver Downtown May 24–26, 2027 [E]. **Atlanta**, where Setup® runs an agency mastermind scoped explicitly to 10–50 FTE marketing agencies, the single tightest firmographic match in the dossier [E].

Second tier, cheaper to reach and less contested by every other AI tool chasing this buyer: **Minneapolis** (Bureau Owner Camp and Digital Summit both landed there in August 2026, and the Bureau format caps at ~20 seats) [E]; **Austin** (PMS February 24–25, 2027) [E]; **Washington DC** (PMS January 28, 2027) [E]; **San Diego and Carlsbad** (Ahrefs Evolve October 12–13, 2026, 600+ practitioners single-track; B2BMX at Omni La Costa February 22–24, 2027) [E]; **Cleveland** (MAICON, October 13–15, 2026) [E]; and **Raleigh–Durham**, **Nashville**, and **Portland** as lower-cost independent-agency markets worth testing [I].

The ideal HQ pattern is **remote-first and distributed, with a nominal metro of record**: agency marketing roles now post as "Remote US" as a default, and the founder's actual working geography is Slack, not a floor plan [E/I]. Two consequences for reach. First, metro targeting is for event presence and geo-fenced paid social, not for office-based prospecting; a 20-person shop registered in Denver may have four people there. Second, and the thing most sellers get backwards: **(Hidden gem)** this buyer's gathering geography is deliberately decoupled from their HQ geography. The rooms where founders actually talk about margin, pricing, and unfunded strategy work are held in Asheville, Stowe, Minneapolis, and Costa Rica, destinations chosen precisely because no client is there, with 20-seat caps and application gates [E]. Chasing the dense metros gets you proximity to their offices and none of their attention. The efficient play is to fund or attend the small remote-destination camps, and to buy reach in the newsletters those founders read from anywhere (Exit Five at 42K, MKT1, Sakas & Company), while reserving the metro list for follow-up density and event logistics [E/I].


### Growth Stage

- **Growth-stage boutique, roughly 15–35 people and $2–5M revenue, 5–10 years in `(most ideal)`.** Promethean's 2026 agency sample averages 31 FTE and $4.43M revenue at 13% after-tax margin, which is the band where the shop has enough delivery staff to run several retainers at once but no one carrying a strategy title. [E] **The condition:** every account now expects positioning and messaging framework work on top of execution, and the unfunded strategic layer lands on the one calendar that also owns new business, the founder's. Margin at 13% makes senior non-billable hours the most visible cost in the P&L. [E][I] **What they buy:** a second opinion that pushes back on a positioning angle inside an hour, for accounts that will never fund a strategist. The agency plan with shared team context matters here because two or three senior people are already improvising the same thinking separately. [E] **Readiness signal:** LinkedIn headcount 15–35, founder still listed as principal, and a Clutch or Semrush partner profile showing retainer-shaped engagements rather than project work. [I]

- **Event-defined: mid-repositioning toward a B2B software niche `(also ideal)`.** The agency is narrowing from a generalist roster to SaaS and tech, usually after a flat year: Promethean puts 2025 average agency revenue growth at 7.5%, with the smallest shops growing slowest. [E] **The condition:** the founder is doing the exact work the product sells, on themselves, with no one senior to argue with. The failure mode is documented in the buying evidence: a founder who had "already done the positioning work," hired a strategist, got exercises and a deck, and six months later had the same clients and the same pricing pressure. [E] *(Source is a rival positioning consultancy's own marketing, directionally useful, not neutral.)* **What they buy:** adversarial pressure-testing of their own narrative and differentiation before it goes on the site, then the same motion applied to client accounts once it works. **(Hidden gem)** The site is rewritten before the client roster changes, so a case-study page freshly re-cut to SaaS logos is a leading indicator of intent: you are catching the founder in the two-month window where they are the buyer *and* the use case. [I]

- **Consolidating agency, 35–50 people with a named strategy lead or first senior hire in the function `(selective)`.** Large enough that strategy has a title attached to it, which changes who evaluates and what they compare against. [I] **The condition:** the tension has shifted from "no capacity" to "quality control across other people's thinking," and the strategy lead is measured on the shop's point of view being distinctive, the person most alert to slop, and most exposed if client work sounds interchangeable. [I] **What they buy:** access to a curated practitioner corpus as a challenge function for a lead who has no peer internally, not capacity relief. [E] **Why selective:** they benchmark against named human authority: Fletch PMM's 2-week positioning sprints, fractional strategists at marketplace-quoted $150–500/hr, where Hivemind has no case study, no review presence, and no quantified outcome to put on the table. [E] Client AI-use clauses also start appearing in their MSAs at this size, and there is no DPA or training-opt-out statement to answer them with. [E][I] Qualify hard on whether the buyer wants argument or artifacts; the product has no export, no template, and no client-facing output mode. [E]
### Market Position

**Position:** These are **specialists**, not leaders or fast-followers, and the specialism is vertical rather than functional: a 10–50 FTE shop sits in roughly the top 12% of a population where 88% of agencies have fewer than 10 FTE, yet it is nowhere near holdco scale (Promethean 2026, n=119: avg 31 FTE, $4.43M revenue, 74% US) `[E]`. That position forces them to compete on category fluency and senior judgment, because they cannot win on media leverage, staffing depth, or price. The squeeze is documented: 7.5% average revenue growth in 2025 with large agencies growing fastest and small slowest, on a 13% after-tax margin against a ~15% long-run norm `[E]`. **(Hidden gem)** The specialist claim is asserted on the website and undefended in the room. One consultant's account of an agency founder captures the failure exactly: "same inability to explain why his agency was different from cheaper alternatives," despite having "already done the positioning work" and having "a deck" `[E]`. The second-order problem nobody prices in: B2B SaaS marketing is among the most heavily written-about verticals on the open internet, so the niche that used to be the moat is the niche a general model imitates most convincingly. Vertical specialization and LLM substitutability are positively correlated, which makes these agencies the *most* exposed specialists, not the least `[I]`.

**Targeting:** Go first at the **squeezed specialist actively competing for new business**, an agency whose case-study page reads B2B software and whose founder is still named on pitches, rather than the generalist fast-follower. (1) **Urgency: 5/10.** The pain is chronic, not acute; the dossier's own read is that "do nothing" is the highest-probability outcome, and the unfunded strategic layer has been absorbed for years without anyone buying `[E]`. Do not price or pace this deal as if a fire is burning. (2) **Buying-path simplicity: 8/10.** Founder-led, no procurement, no committee evidence at this ACV `[E]`. **Cycle-length reconciliation the whole report should inherit:** the brief's 2–4 weeks is the fast end of a plausible 2–6 week reality, so quote 2–4 as the best case and forecast against 2–6 `[E]`. Comparable sub-$5K SaaS bands running roughly five to six weeks is a directional estimate, not a sourced benchmark, and nothing here should rest on it `[I]`. (3) **Budget ownership: 7/10.** The founder signs, and 68% of Promethean's respondents were founders, owners, or partners; capped at 7 because Hivemind publishes no price at any tier, so the "small-tool budget" fit is asserted rather than verified `[E]`. (4) **Observable frequency of the triggering event: 5/10.** The trigger is a competitive pitch or a scope-expansion request ("clients keep asking for more: positioning, GTM synthesis, campaign architecture"), and neither is directly visible from outside `[E]`. Usable proxies only: new case-study publication cadence, Clutch review velocity, an open strategist or new-business role, and Semrush or HubSpot partner-directory activity `[I]`.

**Leverage Strategy:** (1) **Wedge: one live pitch, not a rollout.** Scope the pilot to a single competitive pitch or one at-risk account. The founder brings the angle they already believe and uses the adversarial mechanic to have it argued against before the client does, in the pattern of the vendor's own demo, which answers a request for a launch plan with "Before we talk launch, what's your retention loop?" `[E]` This scope fits the product's real bounds: no ingestion, no export, no integrations, so the input is pasted context and the output is a sharper point of view the founder carries into the room themselves `[E]`. Steer pilot questions toward positioning, narrative, and growth diagnostics rather than SaaS vertical benchmarks, since the documented corpus is web3-scoped and the cross-industry claim is uncorroborated `[E]`. Sell the behavior, not the training data: "trained on real practitioners" is parity language M1-Project already runs in customer voice `[E]`. (2) **Expansion: seats and occasions, not workspaces.** Move from founder-only to the published team or agency plan with shared team context, so the AMs who already "lead strategy" get the same pushback, and let the same thread carry into landing page copy and email sequences so the angle survives the handoff from strategy to delivery `[E]`. There is no per-client memory or client workspace, so expansion means more people and more moments, never more client brains `[E]`. (3) **Compounds: the house method, not accumulated memory.** Displacement gets hard when "every angle gets pressure-tested before it reaches you" becomes a line the founder says to clients in pitches, because ripping the tool out then means retracting a promise `[I]`. What accrues is access to other practitioners' patterns through the curated corpus, not persistence of the agency's own `[E]`. Two honesty constraints to carry into every call: the adversarial mechanic is a vendor claim with zero third-party verification, and there is no white-label mode, so the method is spoken, never handed to the client as a co-branded artifact `[E]`.


### Company Culture And Values

- **They draw a hard line between back-of-house and front-of-house AI, and say so out loud: "Yes but mostly for strategy or operations, never for creative," while the same practitioners call generic model output "a computer that just spits out aggregated slop"** `[E]`. The operating value is that the fingerprint on the deliverable *is* the product; anything that leaves a machine's signature on client-facing work is a liability, not a shortcut `[I]`. Use this by entering through the strategy door only. Sell Hivemind as a second opinion that pushes back before the pitch, never as a draft generator, and let the founder keep authorship of everything the client sees. The execution copy is in scope and should stay in the room, unmentioned in first-touch messaging: leading with it triggers the slop objection before the value lands `[I]`.

- **They ridicule performative adoption in their own peer rooms, saying "the only reason agencies say they are adopting AI is because other agencies are saying," and separately distrust model agreeableness: "Its sycophant mode has been a real problem of late"** `[E]`. The value is intellectual friction: a tool that validates is worth less than nothing to someone whose job is telling clients things they don't want to hear `[I]`. Use this by leading every asset with the product's own demo exchange verbatim. Asked for a full launch plan, it answers "Before we talk launch, what's your retention loop?… Walk me through the first 72 hours." That is the one differentiated claim available, and it is exactly the behavior this audience says the free tools fail at `[E]`.

- **They talk about their own economics in a specific vocabulary: "non-billable," "billable," and the observation that "your agency does not reward non-billable activities," alongside a principal's account that proper brand strategy "requires tens of thousands of dollars in internal resources and tons of time"** `[E]`. The value is that work is judged by whether it can be booked, not by whether it is good; the unfunded strategic layer is a cost center they are already resentful about `[I]`. Use this by pricing and framing the pitch against unpaid pitch hours and routine accounts, not against the flagship engagement. The comparison to make is a founder's speculative Saturday, not a fractional CMO retainer. Note the seller's own "$300/hr" copy anchors against agency expertise, which is the buyer's own margin, and should be cut from anything agency-facing `[E]`.

- **An agency founder who hired a positioning firm for his own client posted the credit publicly: "there's no way we could've gotten here on our own"** `[E]`. The value is that borrowed senior judgment is safe when it is named and unsafe when it is hidden; attribution converts a delegation into a credential `[I]`. Use this by making the practitioner-corpus provenance quotable, and giving founders explicit language for citing it in client conversation. Bound it honestly: the documented corpus is web3-scoped, contributor counts vary across the company's own surfaces, and no third-party verification exists, so the message is "grounded in practitioner playbooks," never a headline number `[E]`.

- **(Hidden gem)** **They brag about the shape of the fee, not just the size. The most-praised agency-founder Slack is "$35 lifetime, one payment, no subscription," endorsed by a well-known agency advisor as "the internet's best no-subscription community for agencies"** `[E]`. The value is fee-structure suspicion, and the second-order cause is that this buyer sells retainers for a living: they know exactly what a recurring line item feels like on the paying end, and they are unusually alert to being on it `[I]`. Use this by removing recurrence anxiety from the first conversation. Offer a single scoped engagement the founder can attach to one live pitch, prove the pushback there, and let expansion follow usage. With no published price, trial, or seat terms anywhere, ambiguity is currently doing the damage a bad price would do, and the fix is a plainly stated entry commitment `[E]`.


### Tools

Docs and knowledge base: **Notion, Google Docs, Google Slides**, signals that the strategy artifact already exists and is already inert. One agency states that Notion is the center of its content operation and every article starts there, with client documentation, SOPs, and briefs in the same workspace [E]. The second-order problem is one founders do not see coming: everything in that workspace was written by them, so any retrieval layer built on top of it, including the custom GPT per client that one agency hand-built [E], can only return the agency's own priors with better formatting, and gets more agreeable the more context you load [I]. That is why the escalation hardens the exact problem it was built to solve. Do not sell storage or recall. Hivemind has no per-client memory and no ingestion at all [E]; sell the second opinion that pushes back. **(Hidden gem)**

Project management: **Asana, ClickUp, Monday.com, Basecamp, Workamajig, Function Point**. Asana appears directly in agency-side job postings [E]; the rest appear in this segment's stack lists [E]. Workamajig and Function Point are built around job costing and billable time [I], so their presence means the founder can already see which hours are recoverable and which are not [I]. The unfunded strategic layer is the work that never gets a job number [I]. That visibility is the wedge: they know what strategy costs them before you tell them.

Communication: **Slack, Slack Connect, Loom, Zoom**. Slack Connect supports shared channels across up to 20 organizations [E], which puts agency and client in one continuous room. That cuts both ways: it is how a founder ends up reading feedback that is clearly from ChatGPT, an inversion practitioners report in first person [E]. It also means anything Hivemind shapes reaches the client with less internal review than a deck would [I]. Sell rigor and challenge, never throughput.

CRM and marketing automation: **HubSpot, Salesforce, Marketo, ActiveCampaign**. HubSpot is named as the hub in an agency's own published B2B SaaS stack [E], and HubSpot Solutions Partner tiering gives you a hard, filterable qualifier for this segment [E]. It is also a warning. This buyer is habituated to tools that write into their stack, and Hivemind has zero integrations of any kind [E]. Position it as a thinking surface adjacent to the stack; any workflow claim will be tested and fail [I].

Analytics and attribution: **GA4, Google Tag Manager, Looker Studio, Mixpanel**, all four surface in live agency and growth-marketing postings [E]. Every one of them measures output after the angle was already chosen [I]. That is the structural reason strategy stays a non-billable absorb: there is no attribution trail for the thinking layer, so nothing to put in front of a client and price against, while every execution channel below it is provable to two decimal places [I]. A founder running a Looker Studio dashboard per client has lived this asymmetry for years without a name for it. Name it on the call. **(Hidden gem)**

Paid media: **Google Ads, Meta Ads, LinkedIn Ads, Microsoft Ads, The Trade Desk, Performance Max, Reddit Ads**. This exact cluster appears as required skills in agency job postings, including The Trade Desk and Reddit Ads [E]. A posting listing a dozen ad platforms describes an agency paid to execute and asked for positioning for free [I]. Reddit Ads is a small tell of a team already testing where B2B buyers actually are [I].

Design and creative: **Figma, Canva, Descript, Loom**, the front-of-house craft layer [E], and the one place this buyer has drawn a hard line. A practitioner states the permission structure verbatim: AI is used "mostly for strategy or operations, never for creative" [E]. That is the single most favorable fact in the stack for Hivemind, whose published scope is positioning, GTM sequencing, and messaging frameworks [E]. Do not stray into asset generation in the pitch; it walks into the slop objection [I].

Web and CRO: **Webflow, Unbounce, Screaming Frog**. Webflow and Unbounce appear in live growth-marketing postings [E]. The messaging framework becomes a homepage within weeks, which is where a weak angle first becomes visible to the client [I]. It is also where the head-to-head lives: Fletch PMM sells two-week positioning-and-homepage sprints against 500+ named B2B software engagements [E], while Hivemind has no case study, named customer, or review anywhere [E]. Compete on latency and challenge, not authority.

SEO, content research, and outbound data, the segment-specific category: **Ahrefs, Semrush, Surfer, Apollo, Clay, Instantly**. One agency publishes Apollo as its outbound default, Clay for enrichment-heavy ABM, and Ahrefs plus Surfer at roughly $300/month combined [E]. That last figure sets the established price band for a serious research tool in this stack [I]. Clay in particular marks the AI-fluent operator building enrichment logic by hand [I], the same person likeliest to test Hivemind against Claude on day one, and likeliest to notice if it simply agrees with them [I].
# Buyer Persona Within Company

### Job Titles

### Primary targets

- Founder [E]
- Co-Founder [E]
- Founder & CEO [I]
- Owner [E]
- Managing Partner [E]
- Partner [I]
- Managing Director [I]
- President [I]
- Principal [E]
- Chief Executive Officer (at 30–50 FTE agencies, where the founder has formalized the title but still runs new-business pitches) [I]
- Chief Strategy Officer (adjacent budget owner: the title that exists only once strategy has been made someone's job, and the person who feels the unfunded strategic layer daily) [I]
- Head of Strategy (adjacent budget owner; at this headcount typically reports straight to the founder and can get a small-tool spend approved in one Slack message) [I]
- Chief Growth Officer (adjacent budget owner; owns new business, so owns the pitch where "why are you different from a cheaper alternative" gets asked) [I]
- Director of Strategy [I]

### Disqualifying or conditional titles

- Executive Creative Director (switch pitch: this is the "aggregated slop" constituency. Lead with the sanctioned use, strategy and diagnosis, never creative output; craft-facing framing loses the room) [E]
- Chief Technology Officer / Head of AI (deprioritize or expect a long evaluation: this account has likely already hand-built custom GPTs per client, and will interrogate the RAG-over-corpus architecture, the gated API, and the absence of any per-client workspace) [E]
- Fractional CMO (conditional: this title is the alternative the product is priced against, not the buyer. If it appears as an outside contractor on the account, the strategic gap is already filled at $4k–20k/mo) [E]
- Vice President of Marketing / Product Marketing Manager (disqualifying for this segment: signals an in-house B2B SaaS marketing org, not an agency. Route to the secondary persona play, where Ignition, Octave and Averi are the live comparison set) [E]
- Director of Client Services / Account Director (deprioritize as entry point: owns delivery throughput and non-billable time, so reads a new tool as another system to maintain rather than a second opinion; useful as a champion only after the founder is sold) [I]
- Chief Financial Officer / Director of Operations (conditional: rare at 10–50 FTE, but where present, expect the client-confidentiality question: where does uploaded client material go, and there is no SOC 2, DPA, or training-opt-out statement to answer it with) [E]

### Responsibilities & KPIs

### Responsibilities

- **New business won: the agency arrives at a pitch with a point of view, not a discovery questionnaire.** This is the responsibility the founder cannot hand off, because the prospect bought the founder. The failure mode is documented in agency-consulting casework: showing up with "the same inability to explain why his agency was different from cheaper alternatives." Hivemind's in-scope work here is pre-pitch pressure-testing of the positioning angle, not the deck. `[E]`

- **Retainer renewal and scope defense: existing clients keep paying the current rate for the current scope.** Renewal is where the slop risk gets priced. A client who has started sending feedback that is "clearly from ChatGPT" is a client already testing whether the strategic layer is worth what it costs. `[E]`

- **Gross margin per account: senior hours stay proportionate to the fee.** At the segment's average 13% after-tax margin, a single account where the founder absorbs a week of unfunded strategic work is the difference between a profitable retainer and a break-even one. `[E]`

- **(Informal) The quality bar on anything strategic that leaves the building.** No job description assigns this and no one measures it, but the founder is the last reader before a positioning doc or campaign narrative reaches a client, and every miss lands on their name. This is the specific reason delegation doesn't relieve the load: one live agency posting advertises that "our AMs lead strategy," so the work is delegated while the accountability is not. `[E]`

- **(Informal) Being the house's second opinion on demand.** Account leads and strategists bring half-formed angles to the founder for the argument, not the approval, and the queue is unscheduled and uncapped. It is judged only in the negative, when the founder is unavailable and weak work ships. This is the function Hivemind's adversarial mechanic is aimed at, and the one that never appears on a P&L. `[I]`
### KPIs

- **Pitch-to-win conversion rate on new business, measured as wins ÷ qualified pitches per quarter.** This is the number the founder quotes to themselves, and it's the one that moves fastest when the opening point of view sharpens. The seller's ROI case lives or dies here, and it currently has no proof: zero Hivemind case studies, named customers, or quantified outcomes exist at any source tier, so this must be instrumented as first-party data from the first ten accounts. `[E]`

- **Non-billable senior hours per month, and their share of founder capacity.** The segment's own vocabulary is billable versus non-billable, and agency-consulting content addressed to these owners states plainly that "your agency does not reward non-billable activities." A defensible claim is displacement of unfunded pressure-testing hours, not a per-brief time saving. Note the dossier's warning: the circulating "45-60 minutes versus 3-4 hours" agency stat is unattributed and must not be used. `[E]`

- **Net revenue retention across the retainer book, plus the count of accounts where strategy is a named, priced line item rather than absorbed scope.** The second half is the one that converts the unfunded strategic layer into revenue, and it is the strongest available ROI framing at this price point given no published Hivemind pricing exists to anchor against. Benchmark context: 7.5% average agency revenue growth in 2025, with small agencies growing slowest. `[E]`


### Buying Committee Role

- **Initiates after a pitch or a scope conversation where the strategic layer showed up unfunded:** a prospect asks for positioning and GTM sequencing on top of the execution retainer, or the founder loses a pitch unable to explain why their shop is different from a cheaper alternative. Internally they do not frame it as buying an AI tool; they frame it as margin. Senior unbillable hours coming out of the founder's week in a business running roughly 13% after-tax, with clients who "opt to skip strategic planning in favor of immediate implementation." [E]
- **Approves alone, off the company card, no procurement and no committee at this deal band.** The person doing the strategic thinking and the person signing the invoice are the same person, which is what makes the 2-4 week cycle plausible. The ceiling is behavioral, not policy: past roughly $500/month the purchase stops being compared to a software line item and starts being compared to a fractional CMO retainer ($4k-20k/mo, marketplace-quoted), a comparison Hivemind loses because it cannot sit in the client room. Note the seller-side complication: no price is published at any tier, so this founder cannot self-qualify and must be quoted live. [I]
- **Personally runs the parallel test before anything advances:** same prompt into the free ChatGPT tab already open, and if the answers rhyme, the evaluation is over. The specific thing they are testing for is whether it disagrees. This buyer distrusts model agreeableness ("its sycophant mode has been a real problem of late"), and the only differentiated proof available is the adversarial move on the seller's demo, where a request for a full launch plan is answered with "what's your retention loop?" Expect the corpus question next, because the documented knowledge base is web3-scoped and a SaaS-focused founder will find that within two queries. `[E]`
- **Pulls in the senior strategist or account lead who will actually use it, and needs them to see it as backup rather than replacement.** The seller's own "without the agency markup" framing lands badly in that room and should be dropped from the demo. **(Hidden gem)** The committee member nobody names is the founder's *client*: agencies report client AI policies now arriving with clauses the agency is asked to sign, so the blocking question is rarely a security review and almost always a single email asking where client discovery material goes and whether it trains a model. Hivemind has no DPA, no SOC 2, and no training-opt-out statement published, which means this arrives as an unanswerable email rather than a negotiation. `[E]`
- **Defines trial success as one moment, not a metric:** a positioning or GTM conversation where it surfaced an angle they had not considered, or told them they were wrong in a way that held up in front of the client. Volume and speed claims do not count and will be read as slop risk. Rollout widens to account leads only after the founder has used an output in live client work without flagging it as AI-assisted, and the real churn risk is not competitive loss but non-adoption after signup. At this ACV the constraint is activation, not evaluation, so the seller should treat the first week's usage as the renewal event. [I]
### Decision Making

### 1) Slop detection
**Impact:** They evaluate output the way they evaluate a junior's draft, scanning for the tells they already recognize. One agency owner puts it this way: "I can spot AI-generated copy from a mile away. And I hate it… It's all starting to sound the exact same." [E] What stalls them is polished, hedged, comprehensive-sounding output. What accelerates them is the product refusing the brief and asking a harder question back, which is exactly the demo asset the company already has ("Before we talk launch, what's your retention loop?"). [E] Open the demo with a deliberately bad prompt from their own world and let it get argued with, rather than showing a clean deliverable. [I]

### 2) Free-tool benchmark
**Impact:** They will run your prompt through ChatGPT or Claude in the next browser tab, because that is the comparison the whole category has trained them to make. A paying reviewer in an adjacent category walked over precisely this ("didn't provide any significant advantage over free alternatives like GPT or Gemini"). [E] Do not defend against the comparison, stage it: same prompt, both windows, and let the difference be the pushback and the practitioner grounding rather than the prose quality. [I] Claims of being "not generic AI, trained on real practitioners" will not carry the moment on their own, since M1-Project already runs that line in customer voice. [E]

### 3) Client-room defensibility
**Impact:** Before capability, they run a political test: could I say out loud, in a client room or on a pitch, that this was involved. This is why the sanctioned use they describe is "mostly for strategy or operations, never for creative." Back-of-house thinking is defensible, front-of-house output is exposure. [E] Sell it as a second opinion that pressure-tests their positioning before the client sees anything, never as something that produces the client-facing artifact, and be ready for the single blocking email about where client discovery material goes. There is no SOC 2, DPA, or training-opt-out statement to point at, so the founder-led answer has to be direct and personal. [E]

### 4) Non-billable math
**Impact:** They price the tool against unfunded senior hours, not against a software line item, because doing strategy properly costs "tens of thousands of dollars in internal resources and tons of time" while the agency "does not reward non-billable activities." [E] The number that moves them is the pitch they can now enter with a point of view instead of questions, or the routine account that finally gets strategic attention it could never justify. [I] Avoid the "replaces the $300/hr strategist" anchor entirely. It is parity language every consultancy uses, and aimed at this buyer it argues against the expertise they bill for. [E]

### 5) Quiet re-decision *(Hidden gem)*
**Impact:** At this deal size the risk is not losing the evaluation, it is losing the account in week three. The founder signs quickly with no committee, then stops opening the tab, and nothing about that shows up as an objection you can answer. [I] The cause is structural: the product has no integrations, no export, and no client workspace, so it lives outside Notion, Slack, and Docs, where the work actually sits, and must be remembered into. [E] Build the close around a specific recurring moment rather than a use case, and treat the first three real pitches or positioning arguments as the activation milestone you personally chase. [I]


### Beliefs

- **"The strategy is the only thing a client can't get cheaper somewhere else. The execution is already a commodity."** They will pay for anything that sharpens the argument they make in a pitch room, and they will not pay for another tool that produces more output faster. (shaped by watching production work get repriced against near-zero-cost AI competitors, "you can't win a price war with someone whose costs are near zero," while their own pitches keep coming down to whether they can explain what makes the agency different from a cheaper alternative) `[E]`

- **"If I can spot AI copy from a mile away, so can the client."** This is the belief that gates every purchase in the category: the tool has to make their thinking better without leaving fingerprints on anything the client sees. It is why they sanction AI back-of-house and refuse it front-of-house, "yes but mostly for strategy or operations, never for creative." (shaped by the daily experience of reading inbound work that all sounds the same, and by "slop" becoming a word this audience uses unprompted) `[E]`

- **"Anything that agrees with me is worthless."** They have decided that the failure mode of general AI is not error but agreeableness, "its sycophant mode has been a real problem of late," and a strategy tool that validates the angle they walked in with is a tool that cost them a client. This is the belief Hivemind's adversarial mechanic is built for, and the only one working clearly in the seller's favor. (shaped by repeated sessions where the model reflected their own brief back to them, and by the buy-side version of the same complaint: agencies that "charge top dollar to echo chamber your brief back at you") `[E]`

- **"ChatGPT is already good enough for this, and it's already open."** The hardest belief to move, and it is not stated as skepticism about AI. They use it constantly. It is a judgment that the strategic layer of an LLM is undifferentiated and that paying for it twice is a tax. Any Hivemind pitch that leads with corpus provenance meets this belief head-on and loses, because it sounds like every other vendor's claim. The dossier's proxy-category verbatim is the exact objection in the wild: no "significant advantage over free alternatives like GPT or Gemini." *(This is the belief the seller must overcome. The demo transcript, being asked about the retention loop instead of handed a launch plan, is the only asset in the dossier that argues against it, and it argues by demonstration rather than by claim.)* (shaped by a year of AI vendors making identical promises, "context never dies," context engines, context layers, brand kits, until the category's language stopped carrying information) `[E]`

- **"Non-billable hours are a business problem I solve by working later, not by buying software."** They do not believe a subscription converts into recovered senior time, because they have bought tools on that promise before and the hours came back anyway. This is why "do nothing" is the most likely outcome: the pain is chronic, the invoice is monthly, and nothing breaks this quarter if they ignore it. (shaped by running at roughly 13% after-tax margin in a shop where the comp structure "does not reward non-billable activities," and by the pattern where clients "opt to skip strategic planning in favor of immediate implementation," so the strategy work has no line item to charge against in the first place) `[E]` `[I]`
### Motivations

### Real Motivation

- **They want to stop being the only person in the shop who can hold a positioning argument.** Every account where the client asks "so what's our angle?" routes back to one calendar. The felt experience is not overwork, it is the low-grade resentment of reading a strategy deck at 11pm that a senior AM drafted and getting to the third slide before realizing they have to rewrite the premise. What they want is not more hours; it is a second opinion that pushes back before they have to be the one who does it. `[I]`
- **They are afraid of being outed as undifferentiated in a room they used to own.** The wound is specific and commercial: a founder unable to explain "why his agency was different from cheaper alternatives," despite having already done the positioning work and having a deck to prove it. The dread peaks in the pitch Q&A, not the pitch. It's the moment a prospect asks what makes this approach yours and the answer comes out sounding like the last three agencies they heard. `[E]`
- **They want AI in the building without their fingerprints on the slop.** This founder uses ChatGPT daily and can spot its output "from a mile away. And I hate it… It's all starting to sound the exact same." The private calculation is that AI is already unavoidable and the only question is whether it touches work a client will ever read. Their own audience has drawn the line for them: sanctioned "mostly for strategy or operations, never for creative." The motivation is permission to use it back-of-house at full force, guilt-free. `[E]`
- **(Hidden gem) They are buying an argument they can attribute to someone other than themselves.** In a founder-led shop, when the founder disagrees with a client's positioning instinct, the disagreement is personal and the client hears ego. When the pushback comes from a corpus of practitioners who have shipped campaigns, the founder gets to be the one *agreeing with the challenge* rather than issuing it. This is the same move an agency founder made publicly after hiring a positioning consultancy: "there's no way we could've gotten here on our own," crediting the outside expert rather than defending his own judgment. The product's adversarial mechanic is more than a thinking aid; it is cover in a client conversation. `[I]` *(Vendor-claimed mechanic, unverified by any customer; the attribution dynamic is our inference from an attested quote.)*

### Stated Motivation

- **"We're bleeding senior hours into non-billable strategy."** Defensible, quantifiable, and the one framing that survives a partner conversation. It sits on real economics: doing brand strategy properly costs "tens of thousands of dollars in internal resources and tons of time," while the agency "does not reward non-billable activities," against an industry average 13% after-tax margin (Promethean 2026, n=119 agency owners, 74% US). `[E]`
- **"We need to show up to pitches with a point of view instead of questions."** The efficiency version of the status fear. It reframes strategy tooling as new-business investment rather than overhead, which makes it approvable at a routine account's budget rather than a flagship's. `[E]`
- **"It's cheaper than the fractional strategist and faster than the sprint."** They will name the comparison set out loud: fractional CMOs quoted at $150–500/hr and $4k–20k/mo retainers, positioning sprints from named consultancies. This is the safest thing to say and the least true to what actually drives them: nobody has ever renewed a tool because it was cheaper than a person they were never going to hire. Treat it as the sanctioned reason, not the buying reason. `[I]` *(Rate bands are T2 marketplace pages with no stated n.)*

### Why Now

- **A client sent back feedback that was obviously AI-generated.** One practitioner described exactly this: "I just got feedback from a client that is clearly from ChatGPT." The quarter this happens is the quarter the founder stops asking whether AI threatens their delivery and starts asking what they can charge for that a compliant model will never produce for the client directly. The scramble is defensive and it has a deadline: the next scope conversation on that account. `[E]`
- **A retainer renewal lands on an account where the strategy deck from the last engagement never moved anything.** The pattern is documented in a consultant's account of a founder who ran a full positioning engagement, exercises, ICPs, value props, a deck, a tagline, and six months later had the same clients and the same pricing pressure. When that account comes up for renewal and the founder has to justify the strategic line item, the search starts. `[E]` *(Source is a rival positioning consultancy's own marketing, the narrative is real practitioner-facing content, not neutral research.)*
### Key Insights

### 1) The Client's ChatGPT

**Insight:** The agency founder is defending the wrong flank. They are braced for the embarrassment of being caught using AI, but the live wound recorded in practitioner forums is the inverse: "I just got feedback from a client that is clearly from ChatGPT." When the client can generate a second opinion for free, the founder's strategic judgment gets repriced from the buy side, and no amount of restraint on the agency's part changes that. [E]

**Impact:** → Action: Open discovery with "When's the last time a client sent you feedback that was obviously ChatGPT?" Then position Hivemind as the thing that lets them walk into the next call with a counter-argument the client's model would never volunteer. The site's own demo line ("Before we talk launch, what's your retention loop?") is the artifact to show. Do not sell hours saved; sell the ability to be less agreeable than the client's AI. [I]

### 2) Non-Billable First

**Insight:** This founder does not evaluate tools on capability, they evaluate on which ledger the tool moves work between. Strategy is unfunded: clients "opt to skip strategic planning in favor of immediate implementation," the internal cost runs to "tens of thousands of dollars in internal resources," and agency comp structures "do not reward non-billable activities." A tool that makes billable work faster is a nice-to-have. A tool that makes the unfunded strategic layer survivable is a margin decision at 13% average after-tax margin. [E]

**Impact:** → Action: Price and pitch against the pitch deck, not against the retainer. First-call question: "How many hours went into your last new-business pitch that you didn't bill?" Frame the subscription as the cost of one unwon pitch, and target the pre-sale moment, speculative positioning work for a prospect, because that is the hours block with no client to charge. [I]

### 3) Strategy, Never Craft

**Insight:** The permission structure is already written and it is narrower than sellers assume: this audience sanctions AI "mostly for strategy or operations, never for creative." That is not squeamishness, it is a defended boundary. The craft is what they believe the client is paying a human for, and the strategy layer is where they will accept machine help without feeling they have sold the shop. Hivemind's stated scope includes landing page copy, email sequences, and social content, which crosses that line. [E]

**Impact:** → Action: Lead every asset with pressure-testing positioning and never with copy generation, even though the product does both. Demo the challenge behavior first and let the execution copy surface only as an unprompted second beat once the founder asks. On the site and in outbound, cut "end to end": it is on the parity list and it reads as an offer to automate the craft they are protecting. [I]

### 4) The Notion Incumbent

**Insight:** The real competitor is not ChatGPT, it is the founder's own Notion workspace: "Notion is the center of our content operation… Every article starts in Notion," plus the escalation path where an agency builds a custom GPT loaded with a client's context. That workaround is functionally a mirror: it retrieves what the agency already believes and returns it in cleaner prose. It cannot disagree, because it was built from their own material, and the founder has not noticed that this is the defect rather than the feature. [E]

**Impact:** → Action: Name the mirror on the first call: "Your custom GPT is trained on your own decks. What has it ever told you that you didn't already think?" Then run one live query against a positioning angle the founder currently believes, and show the corpus surfacing a practitioner counter-pattern from outside their book. Position against retrieval-of-self, not against generic AI, and never claim per-client memory, because that is the axis where Notion wins and the product has nothing shipped. [I]

### 5) Sold By Attribution

**Insight:** This buyer's fear of looking less expert is resolvable, and one of them has already published the resolution: an agency founder who hired a positioning consultancy for a client posted "there's no way we could've gotten here on our own," crediting the outside expert publicly and gaining status rather than losing it. The reputational danger is not using outside strategic input, it is being caught concealing it. Hivemind currently sells the opposite reflex with "without the agency markup" and "stop paying $300/hr," which frames it as a thing to hide from clients rather than cite in front of them. [E]

**Impact:** → Action: Kill the agency-markup line from anything an agency founder will see; it argues against the fee they bill. Replace it with a citable frame: the corpus is practitioner pattern recognition from outside their client book, the kind of input a founder names in a pitch. Then ask on the call: "Would you be comfortable telling a client you pressure-tested this against a network of operators?" A yes is a qualified deal; a no means you are selling a secret, and secrets churn silently. [I]
# Jobs To Be Done

### 1) Walk into a new-business pitch already holding a point of view about the prospect's market

**Circumstances:** The founder is on the second call with a Series B SaaS prospect who has already talked to two other shops. Everyone arrives with a credentials deck and a discovery questionnaire; the one who arrives with an actual read on the prospect's positioning wins the room. Doing that properly is exactly the work an agency principal described as requiring "tens of thousands of dollars in internal resources and tons of time," which is unaffordable on a deal that may not close. So the founder either speculates thinly or asks questions and hopes competence reads as rigor. Losing here costs the whole retainer, not a line item. [E]
**Importance:** 9/10, supported by verbatim agency-principal language on both the value of arriving with insight and the cost that prevents it. [E]
**Currently served by:** Personal experience plus a fast skim of the prospect's site and competitors, sometimes a ChatGPT pass the night before; the more disciplined shops buy a positioning sprint from Fletch PMM for real clients but never for prospects. **PARTIALLY.** [E]
**Frequency:** Every live pitch, so weekly to monthly depending on pipeline (spikes in Q4 and Q1 when annual budgets reset). [I]
**Moment:** Right after the intro call ends and the founder has four days, no budget, and a prospect who will compare three points of view side by side. [I]

### 2) Produce a real strategic layer for a routine account without spending the founder's own week on it

**Circumstances:** A mid-size retainer client asks for positioning work on a new product line. The account is profitable at 13% after-tax margin only if senior hours stay scarce, and the agency's comp structure "does not reward non-billable activities," so nobody but the founder will absorb it. Clients themselves "opt to skip strategic planning in favor of immediate implementation," which means the work is expected but not separately funded. The founder ends up doing it in evenings, or the account gets execution with a thin strategic story bolted on. The second outcome is how retainers quietly become commodity work. [E]
**Importance:** 9/10, two independent verbatim sources establish both that clients skip funding strategy and that agencies structurally punish the hours it takes. [E]
**Currently served by:** Founder's unbilled time, recycled frameworks from a previous client, or a Notion-based process where "every article starts in Notion" but the strategy behind it does not exist as a maintained artifact. **UNSERVED.** [E]
**Frequency:** Continuous across the book, with an acute spike at every account's quarterly planning and any client-side product launch. [I]
**Moment:** When the founder opens their calendar for the coming week and sees three client strategy asks with no billable hours attached to any of them. [I]

### 3) Have someone competent argue back before the client does

**Circumstances:** The founder has landed on an angle for a client's repositioning and believes it is right. There is no peer in the building senior enough to tell them it is wrong, and the account team's incentive is to agree and start producing. The next honest reaction arrives from the client in a room with six people in it. This audience already distrusts agreeable AI, describing sycophancy in current tools as "a real problem," so a compliant assistant does not fill the gap. The cost of finding out late is a rework cycle plus visible loss of authority. [I]
**Importance:** 7/10, inference; the sycophancy complaint is verbatim but no source shows a founder describing this as a job they're hiring for. [I]
**Currently served by:** Calling a friend who runs another agency, posting a sanitized version in a paid mastermind like Setup® or Grow Your Agency, or waiting for the client. **UNSERVED** at the speed the work actually moves. [E]
**Frequency:** Several times a month; spikes on any recommendation the founder feels unusually confident about. [I]
**Moment:** When the deck is 80% built, the angle feels right, and there is no one left to check it against before it becomes a client-facing commitment. [I]

### 4) Be able to say, out loud and quickly, why this agency is worth more than the cheaper option

**Circumstances:** A prospect or an existing client names a competitor at half the fee. The founder has a positioning deck somewhere from a prior exercise and still fumbles the answer. One consultant's case notes describe exactly this: a founder who insisted he had "already done the positioning work," who "had a deck," and who showed the "same inability to explain why his agency was different from cheaper alternatives." The failure is not intellectual, it is that the differentiation was never stress-tested against a live objection. Each fumble costs either the account or the rate. [E]
**Importance:** 9/10, near-verbatim description of this exact failure from a named consultant's account of an agency founder. [E]
**Currently served by:** A dormant positioning deck, a website About page, and improvisation in the moment. **PARTIALLY**, and the served part decays fast. [E]
**Frequency:** Every renewal conversation and every competitive pitch, so monthly at minimum. [I]
**Moment:** When a client says a competitor quoted less and the founder has roughly eight seconds to answer without sounding defensive. [I]
### 5) Make sure the work does not sound like everyone else's

**Circumstances:** The founder reviews a messaging framework their team produced and it reads correct and completely interchangeable. They know the tell because they have it themselves: "I can spot AI-generated copy from a mile away. And I hate it… It's all starting to sound the exact same." Practitioners in this segment call the failure mode plainly, output that is "weak and fluffy," or "a computer that just spits out aggregated slop." The exposure is not error, it is sameness, and sameness is what a client points at when arguing the work could have come from anywhere. This is why the audience sanctions AI "mostly for strategy or operations, never for creative." [E]
**Importance:** 9/10, multiple first-person practitioner verbatims naming the fear and its exact vocabulary. [E]
**Currently served by:** Founder review as the last line of defense, plus informal bans on AI in client-facing craft work. **PARTIALLY** (it catches slop but does not generate a distinctive angle to replace it). [E]
**Frequency:** Every deliverable review; spikes on anything the client will see unedited. [I]
**Moment:** When the founder reads a draft, feels nothing land, and cannot name what is missing beyond "this could be for anyone." [I]

### 6) Hold the fee when the client assumes AI made the work cheap to produce

**Circumstances:** The client knows the agency uses AI, because everyone does, and reasons that costs fell. A cost-plus conversation starts where a value conversation used to be. The counter-frame circulating in the agency community is that nobody expected a discount because a photographer used autofocus, and that discounting AI-assisted work "trains clients to devalue it," but the founder needs an argument grounded in something other than analogy. The structural warning from an agency principal is blunt: "you can't win a price war with someone whose costs are near zero." Losing this argument once resets the rate card for the relationship. [E]
**Importance:** 7/10, the price-war line and the discount counter-frame are attested, but the framing is agency-authored opinion, not measured buyer behavior. [E]
**Currently served by:** Ad hoc justification in the moment, value-based-pricing content from Blair Enns and 2Bobs, and mastermind threads. **UNSERVED** as a repeatable argument tied to a specific engagement. [I]
**Frequency:** At every scope and renewal negotiation; rising steadily. [I]
**Moment:** When a client asks how long the work actually took, and the honest answer would cost the founder the rate. [I]

### 7) Stay the senior voice in the room now that the client has their own AI

**Circumstances:** The inversion has already happened and it is documented in first person: "I just got feedback from a client that is clearly from ChatGPT." The client now generates critiques, alternative angles, and counter-copy between meetings, and arrives with them. The founder's response cannot be more output, because output is the thing the client just proved they can get for free. What survives is judgment the client's compliant model will never volunteer: the reason a recommendation is wrong, not another version of it. Failing to hold this ground turns the agency into a production vendor inside one budget cycle. [E]
**Importance:** 8/10, the triggering event is verbatim and dated; the reframing of what the founder must do about it is analyst inference. [E]
**Currently served by:** Improvised in-meeting defense; one documented thread escalates to running the client's AI feedback back through AI, and to staging anti-AI demonstrations. **UNSERVED.** [E]
**Frequency:** Rising month over month across the client book. [I]
**Moment:** When a client forwards a paragraph of AI-generated feedback and expects the agency to respond to it as if it were considered judgment. [I]

### 8) Choose one angle to commit to and kill the other four fast

**Circumstances:** Discovery produced five plausible positioning directions and each has an internal advocate. The founder is the tiebreaker, and there is no data yet because nothing has shipped. Message testing exists as a discipline. Wynter returns verified-buyer reaction in under 48 hours, priced per test, but it is rarely funded at boutique scale for a mid-size account. So the choice gets made on the founder's instinct, then defended for the next six months. Choosing wrong is not usually caught until renewal, which makes it the most expensive quiet failure in the book. [E]
**Importance:** 6/10, inferred as a distinct job; the dossier evidences the alternatives and the price of testing, not founders describing the decision as a hired job. [I]
**Currently served by:** Founder instinct, borrowed frameworks from Product Marketing Alliance and similar, and workshops run to manufacture consensus rather than resolve merit. **PARTIALLY.** [E]
**Frequency:** Once per major client engagement, roughly quarterly. [I]
**Moment:** When the workshop ends with five directions on the wall and everyone turns to the founder. [I]
### 9) (Hidden gem) Raise the floor on strategy the founder did not personally do

**Circumstances:** Delegation of strategy already happens in shops this size. One agency's live posting advertises "our AMs lead strategy" as the reason to join. The constraint is not whether an account manager can run it, it is that the founder cannot review every framework at the depth that protects the agency's reputation, so quality varies by who happened to own the account. The founder finds out about a weak strategic recommendation when a client questions it, not when it is written. What is missing is a standard the AM can check their own thinking against before it reaches either the founder or the client. Every unreviewed deliverable is an unpriced reputational bet. [E]
**Importance:** 7/10. The delegation is evidenced by a real job posting; the quality-variance consequence is inference. [E]
**Currently served by:** Founder review when time allows, plus reused templates from prior accounts. The audience's stated preference is "don't reinvent the wheel." **UNSERVED** as a consistency mechanism. [E]
**Frequency:** Every deliverable produced without founder involvement, which is most of them past 15 employees. [I]
**Moment:** When the founder skims a framework an hour before the client call and has to decide whether to fix it or let it go. [I]

### 10) (Hidden gem) Attribute the uncomfortable recommendation to something other than personal opinion

**Circumstances:** The founder needs to tell a client their whole category narrative is wrong, or tell their own team that six weeks of work should be scrapped. Delivered as personal judgment, it reads as taste, and taste is arguable by anyone in the room. Delivered as pattern, this is what happens to companies that sequence a launch this way, it reads as evidence and ends the debate. This is why a named authority carries a premium: an agency founder who bought a positioning sprint publicly credited the outside expert with "there's no way we could've gotten here on our own," converting an unpopular conclusion into a defensible one. Without external grounding, the founder must spend relationship capital every time they are right. [E]
**Importance:** 7/10. The buyer's own behavior of publicly crediting outside expertise is verbatim; reading it as a cover-seeking job is inference. [E]
**Currently served by:** Hiring a named consultancy for the accounts that can fund it, or citing April Dunford and Fletch frameworks secondhand. **UNSERVED** for the 80% of engagements that cannot fund an outside name. [E]
**Frequency:** Several times a quarter; spikes whenever the founder's honest read contradicts what the client already believes. [I]
**Moment:** When the founder knows the right answer, knows the client will resist it, and has nothing to point at except themselves. [I]


# Buyer's Journey

### 1) Trigger, no fixed duration: the pain is chronic, so the clock starts on a specific event, not a build-up [I]

**What happens:** The founder hits one of three events: a pitch where they arrived with questions while a competitor arrived with a point of view, a client forwarding feedback that is "clearly from ChatGPT," or a week where three client strategy asks landed with no billable hours attached to any of them. Nobody else joins yet. The internal artifact is not a document, it is a note-to-self or a Slack message to one peer founder, usually phrased as a complaint rather than a requirement. [E]
**Information sources:** Peer conversation inside a paid agency community (Grow Your Agency Slack, Setup® Agency Mastermind forums of 6-8 non-competing agencies); the Glassdoor Advertising bowl, where practitioners post anonymously about exactly these events; 2Bobs or Build a Better Agency in the car. [E]
**Exit criteria:** "This is going to keep happening, and I'm the only one who can fix it." Movement requires the founder to reattribute the problem from bad luck to structure. [I]
**Seller's move:** Publish against the trigger event, not the product. The one asset that works here is the adversarial demo verbatim, where a request for a full launch plan is answered with "Before we talk launch, what's your retention loop?" It names the missing behavior before it names a tool. [E]

### 2) Self-education, 3 to 10 days, discontinuous, done in gaps between client work [I]

**What happens:** The founder searches at the artifact level rather than the category level, because no settled category name exists and G2's nearest 2026 addition is "AI marketing agents," which describes campaign automation and misdescribes this product. They type things like "AI positioning tool," "ICP generator," "AI marketing brief tool." Still solo. The artifact produced is a browser-tab pile and, in disciplined shops, three names in a Notion page. [E]
**Information sources:** ChatGPT and Claude asked directly for options, noting that G2 now pipes verified reviews and buyer intent into both, so the shortlist is partly assembled by the incumbent alternative; Exit Five and Product Marketing Alliance threads; LinkedIn posts from Fletch PMM, April Dunford, and Emily Kramer. [E]
**Exit criteria:** "There's a real category of thing here, and it isn't just a wrapper." The founder must find at least one credible claim they cannot get from the tab already open. [I]
**Seller's move:** Contest the artifact-level queries with proof of the mechanic, not the corpus. Corpus-depth claims are parity and the documented corpus is web3-scoped, so lead with the operating rule that grounds every answer in practitioner material before advising. [E]
### 3) Shortlist, 2 to 5 days, short because the free alternative is the default and only two or three names survive [I]

**What happens:** The list collapses to Hivemind, M1-Project or a similar generation tool, and "keep using ChatGPT." A senior AM or the head of strategy may be shown the tabs, but this is consultation, not a committee. The artifact is a comparison the founder does in their head against one question: which of these will tell me I'm wrong. [I]
**Information sources:** Vendor sites read side by side, where the founder encounters the parity wall directly. Averi's "context never dies," Octave's context engine, M1-Project's ChatGPT contrast in customer voice; Product Hunt for launch signal; a G2 or Capterra search that returns nothing for Hivemind. [E]
**Exit criteria:** "One of these does something the others don't, and I can say what it is in a sentence." Sameness at this stage routes straight back to the free option. [E]
**Seller's move:** Give the founder that sentence and make it the pushback claim, since it is the only capability the parity list does not already contain. Never counter on memory or compounding context. There is no per-client memory to defend, and four competitors own the language. [E]

### 4) Evaluation, 1 to 2 weeks: the whole cycle runs 2 to 6 weeks, faster at the fast end than the brief's 2-4 assumes [E]
**What happens:** The founder tests it on live work, not a sandbox, usually a real account's positioning problem where they already hold a private opinion. They are grading one thing: whether it argues, or whether it agrees. This audience already names sycophancy in current tools as "a real problem," so agreeableness is a fail condition, not a nice-to-have. The artifact is a chat transcript they screenshot and paste to one other person. [E]
**Information sources:** The tool itself, in a single sitting. The company states most users get actionable insight in the first conversation with no onboarding; a sanity-check post in a mastermind Slack; a peer founder's opinion on whether this is "actually different." [E]
**Exit criteria:** "It told me something I hadn't considered, and it didn't sound like everything else." Fluffy or generic output ends the evaluation immediately, and the audience's own word for that outcome is slop. [E]
**Seller's move:** Route the trial toward a live account problem rather than a demo scenario, and score the session on challenge quality. **(Hidden gem)** Insist the first test be run on a *prospect's* market rather than an existing client's, because the pitch use case carries no client-confidentiality question and therefore skips the blocking objection in the next stage entirely. [I]

### 5) Decision & rollout: signature in a day, rollout 30 to 60 days, and this is where the deal is actually won or lost [I]

**What happens:** The founder signs alone, on a card, with no procurement and no security review at this price band. No dollar threshold for triggering review at agencies this size exists in any source, so treat its absence as unstructured rather than confirmed. Then one blocking question arrives, usually by email and usually from an account lead: are we allowed to put this client's material into it. The artifact is either a one-line internal AI-use rule or a quiet decision to keep it founder-only. [E]
**Information sources:** The vendor's own site, checked for a data or training statement and finding none published; the client's own AI policy clauses, which agencies are increasingly asked to agree to; the account lead who raised the question. [E]
**Exit criteria:** "I can answer the client-data question in one sentence, and someone other than me opened it this week." [I]
**Seller's move:** **(Hidden gem)** Treat the client's contract as the unnamed buying committee member and hand the founder a prewritten answer for it, then seed team plans deliberately. Shared team context is the only published multi-user capability, and non-adoption after signup, not competitive loss, is the documented failure mode for deals in this band. [E]
# Goals Or Objectives

### 1) Get paid for the strategic layer instead of absorbing it into founder nights and weekends

**Explanation:** Concretely, this looks like positioning, GTM sequencing, and messaging framework work moving from an unpriced courtesy inside an execution retainer to a named, scoped line item the client approves. Today the economics run the other way: clients "opt to skip strategic planning in favor of immediate implementation," the agency's own comp structure "does not reward non-billable activities," and doing it properly is described by one principal as costing "tens of thousands of dollars in internal resources and tons of time." [E] The founder's ultimate goal is margin that survives the strategy work rather than being eaten by it. [I] Hivemind's role here is narrow and honest: it compresses the hours between "we need a point of view on this account" and "here is a defensible angle," working conversationally across positioning options, GTM sequencing, and messaging frameworks. [E]

**Importance:** Their performance is judged on retainer profitability, and at a 13% average after-tax net margin (Promethean 2026, n=119, avg 31 FTE, avg $4.43M revenue), the senior unbilled hour is the single most expensive thing in the building. [E]

**Impact:**
- Strategy stops being the thing that happens after 7pm and starts appearing on a scope of work with a number attached. [I]
- Accounts that could never justify a strategic engagement get one anyway, because the input cost drops below the threshold where the founder has to ration it. [I]
- The internal argument about whether to write off strategy hours disappears from the monthly P&L review. [I]

**Why Critical:** If strategy stays unfunded, it gets rationed to the two flagship accounts, and every other client relationship quietly degrades into order-taking. Order-takers get compared on rate. [I]
### 2) Stop being the only person in the shop whose strategic judgment the work depends on

**Explanation:** Success looks like an account manager or senior IC producing a positioning angle the founder can sign off on with one round of comment, not rebuild from scratch. Delegation already happens in shops this size. One live agency posting for this segment advertises "our AMs lead strategy," so the constraint is not permission, it is quality control. [E] The founder's real objective is to leave a pitch prep session without having personally authored the thinking in it. [I] Hivemind's contribution is a second opinion that pushes back, available to anyone on the team, not just to the person with twenty years of pattern recognition. [E]

**Importance:** In a founder-led agency the person doing the strategic thinking and the person signing the invoice are the same person, which is why the sale is fast, and why every strategic deliverable is bounded by one calendar. [E]

**Impact:**
- Pitch prep can start without the founder being in the room for hour one. [I]
- The team's first draft arrives closer to defensible, so the founder's edit becomes a sharpening pass rather than a rewrite. [I]
- Vacation, conference travel, and back-to-back client days stop functioning as a hard stop on strategic output. [I]

**Why Critical:** A shop where only the founder can think strategically has a ceiling at the founder's bandwidth and a valuation problem at exit. It also breeds the specific exhaustion of being needed for every deck at 11pm. [I]

### 3) Use AI on the strategic work without the output being recognizable as AI

**Explanation:** The bar here is set by the buyer themselves: an agency owner writing in the first person says "I can spot AI-generated copy from a mile away. And I hate it… It's all starting to sound the exact same." [E] Success is AI in the thinking loop, not on the page. The founder uses it to interrogate an angle, then writes the client-facing artifact in their own voice. This audience already grants itself that permission, sanctioning AI "mostly for strategy or operations, never for creative." [E] Hivemind sits precisely inside that sanctioned zone, and its claimed mechanic is challenge rather than generation: asked for a full launch plan, the published demo answers "Before we talk launch, what's your retention loop?… Walk me through the first 72 hours." [E] Note the vendor-claim status of that mechanic once and carry it: no third-party verification, review, or case study exists. [E]

**Importance:** They are judged on whether the work is credited as the agency's own thinking, and the felt failure is precise: the flush of being asked, in a room, whether a deck was AI-generated. [I]

**Impact:**
- AI moves from the drafting stage to the interrogation stage, which changes what the client ever sees. [I]
- The founder gains a defensible answer to a client's AI-use question, because the tool never touched the deliverable. [I]
- Team members stop pasting client material into whatever consumer chat window is open, which narrows an unmanaged confidentiality surface. [I]

**Why Critical:** One client who spots the tell prices the entire relationship down, and the story travels. In a segment where the product is judgment, being caught outsourcing it is not an efficiency embarrassment. It is a credibility event. [I]

### 4) Have a real answer to "why you and not the cheaper option" before the next pitch

**Explanation:** Concretely: walking into new business with a point of view on the prospect's market rather than a discovery questionnaire, and being able to name what the agency does that the alternative does not. The failure mode is documented in a consultant's case narrative of an agency founder with "the same inability to explain why his agency was different from cheaper alternatives." He had already done the positioning work, and he had a deck. [E] Note the source is a rival positioning consultancy's own marketing; the pattern is credible, the framing is sold. [E] The founder's underlying goal is to stop competing on rate. [I]

**Importance:** New business won is the number their year is judged on, and the differentiation argument is the one moment in the sales cycle where the founder is personally, visibly exposed. [I]

**Impact:**
- Pitches open with a market observation instead of a capabilities slide. [I]
- Speculative pre-sale strategy becomes affordable enough to attempt on more than the one deal worth chasing. [I]
- Rate conversations start later in the cycle, because the comparison set has already been narrowed. [I]

**Why Critical:** Undifferentiated agencies get shortlisted alongside three others and chosen on price. That is a compounding trap: lower rates fund less senior time, less senior time produces flatter thinking, flatter thinking wins on price again. [I]
### 5) (Hidden gem) Sell the one thing a compliant model will never give their client — disagreement

**Explanation:** The founder is braced for the wrong exposure. The threat is not that AI writes their strategy; it is that the client now has AI too. One practitioner reports it plainly: "I just got feedback from a client that is clearly from ChatGPT." [E] That reprices judgment from the buy side, not the sell side, and in a category where a peer warns "you can't win a price war with someone whose costs are near zero," the only durable product left is the pushback a model trained to agree will never give the client. [E] Achieving this goal means the agency's stated deliverable shifts from "we will produce the strategy" to "we will tell you when you are wrong," and the founder buys challenge for themselves rather than automating it away. [I] Hivemind's fit is the mechanic, not the memory: adversarial retrieval over other practitioners' patterns. Their own patterns do not persist here, because the product has no per-client memory. [E]

**Importance:** Their renewal conversations are increasingly judged against what the client could have produced themselves, and every deliverable a client's own model could have generated is a deliverable they will eventually stop paying for. [I]

**Impact:**
- Scope language changes from volume of output to quality of counsel, which changes what a renewal is priced against. [I]
- The founder starts practicing being challenged before the client challenges them, which is a rehearsal loop they currently have no access to. [I]
- Client-supplied AI feedback becomes something the agency engages on the merits instead of a moment of quiet humiliation. [I]

**Why Critical:** Agencies that keep selling production against clients holding near-zero-cost production tools are negotiating their own commoditization one renewal at a time. The end state is not dramatic. It is a slow slide from strategic partner to vendor, noticed only when the retainer is not renewed. [I]


# Problems

### 1) Problem: The strategic layer clients ask for is unbillable, so it gets absorbed into founder hours or skipped entirely

**Description:** Positioning, GTM sequencing and messaging frameworks are the most senior work in the shop and the hardest to put on an invoice. Clients "opt to skip strategic planning in favor of immediate implementation," and the agency's own comp structure "does not reward non-billable activities," so nobody inside is incentivized to do it properly either [E]. Doing it right costs, in one agency principal's own words, "tens of thousands of dollars in internal resources and tons of time," which is why it keeps getting compressed into a weekend rather than resourced [E]. Effort is not the missing ingredient. The unfunded strategic layer persists because the work is real, the demand is real, and the line item is not [I].

**Consequence:** Strategy gets done at the last defensible moment and at the lowest defensible depth, pushing execution start dates and squeezing delivery teams; senior unbillable hours land directly on a 13% average after-tax margin, the thinnest expense line an owner controls (Promethean 2026, n=119) [E]; and the client experiences the agency as an executor with opinions rather than a strategic partner, which is the exact perception that caps retainer size at renewal [I].

**Solution:** Hivemind is priced and packaged as a tool, not an engagement, and covers the company's own enumerated scope (positioning options, go-to-market sequencing, messaging frameworks, competitive differentiation, narrative development) in conversation [E]. That moves the unfunded strategic layer from "senior hours we can't bill" to a fixed monthly cost the founder can absorb across every account, including the routine ones that never justified a strategist [I]. Note the bound: no published price at any tier, so this must be qualified on the call, not assumed [E].

### 2) Problem: In a founder-led shop the strategist and the invoice-signer are the same person, so strategic capacity is bounded by one calendar

**Description:** The founder is still on new-business pitches and still on strategy calls, which is why the sale is fast and uncommitteed and why the bottleneck is structural rather than procedural [E]. Delegation demonstrably happens at this size, one live agency posting advertises "our AMs lead strategy," so the constraint is not that the work cannot be handed off [E]. It is that the founder remains the quality bar, and there is no senior peer in the building to check the angle against before it goes to a client [I]. Hiring one costs $150–500/hr or a $4k–20k monthly retainer, economics that never pencil for a routine account or for speculative pitch work (T2 marketplace pages, no stated n) [E].

**Consequence:** Angles ship unchallenged or wait on the founder's next open block, stalling briefs across multiple accounts at once; pitch work that could have won a retainer goes out at the depth one person could reach by Thursday, and the fractional-strategist alternative stays priced out of everything but the flagship client; internally, the team learns to route every strategic question upward, which entrenches the bottleneck the founder is trying to escape [I].

**Solution:** The product is available at conversation latency with no onboarding step, so a pressure-test happens inside the hour the angle is written rather than at the next calendar opening [E]. Team and agency plans exist with shared team context, letting an AM or senior strategist run the same challenge before it reaches the founder [E]. This is capacity relief, not judgment replacement, and should be sold that way to a buyer whose product is judgment [I].

### 3) Problem: The obvious relief valve is AI, and this founder can already recognize AI output on sight

**Description:** The buyer is not AI-skeptical. They use ChatGPT daily and say so, and in the same breath: "I can spot AI-generated copy from a mile away. And I hate it… It's all starting to sound the exact same" [E]. The felt experience is not caution, it is revulsion at the specific texture of the thing, described by practitioners as "weak and fluffy" and, more bluntly, "aggregated slop" [E]. Compounding it, the seller's own price anchor argues against the buyer's business model, pitching escape from "the agency markup" and "$300/hr" advice to the person charging it [E]. There is one permission structure that cuts in the product's favor: this audience already sanctions AI "mostly for strategy or operations, never for creative" [E].
**Consequence:** The founder rules out an entire tool category on identity grounds before evaluating any of it, keeping the strategic bottleneck intact. The slop risk turns an otherwise routine tool decision into a positioning decision about what the agency is, which is why deals in this band stall rather than lose [I]. And if AI-flavored work does reach a client, the relational cost lands on the one asset the agency actually sells, which trade coverage frames as "expertise, strategic judgement and business outcomes" rather than hours [E].

**Solution:** Sell into the sanctioned lane. Hivemind's claimed mechanic is adversarial rather than generative, it challenges assumptions and surfaces the question that wasn't asked, per its own demo: "Before we talk launch, what's your retention loop?… Walk me through the first 72 hours" [E]. That is back-of-house pressure-testing, not front-of-house craft, which is precisely where this audience has already granted permission [I]. Carry the bound every time: this is a vendor claim with zero third-party verification [E].

### 4) Problem: Every alternative makes the same three claims in the same words, so comparison collapses back to the free thing already open on the desktop

**Description:** "Trained on real practitioners, not generic AI" is not a differentiator. M1-Project runs that exact contrast in customer voice, and brand.ai promises briefs that aren't "generic frameworks" [E]. Persistent context is equally crowded: Averi's "context never dies," Octave's GTM Context Engine, Jasper IQ's context layer, AirOps' Brand Kit [E]. When category language is interchangeable, the founder has no basis to separate paid options and defaults to the incumbent, as a paying reviewer of an adjacent tool put it: no "significant advantage over free alternatives like GPT or Gemini" [E]. Everything sounds the same, so nothing gets bought.

**Consequence:** Evaluation stops at the free tier and the trial never converts, with the dominant outcome being non-purchase rather than competitive loss. Budget that would have gone to a strategic tool stays in seats the shop already pays for. And the founder, having tested three lookalike tools and kept the free one, becomes materially harder to re-approach next quarter [I].

**Solution:** Do not compete on corpus provenance or memory, both of which are parity claims and one of which Hivemind cannot make at all [E]. Compete on the single behavior no competitor's published positioning claims: a tool built to disagree with the founder rather than execute for them. The demonstration is the demo, refusing the requested deliverable and asking the prior question, and it is reproducible in a first conversation with no onboarding, which makes it a sales asset rather than a copy claim [E].

### 5) Problem: The founder is braced for the wrong exposure. The client now has AI too, and judgment is being repriced from the buy side

**Description:** The anticipated threat is producing work that sounds machine-made. The attested event is the inversion: "I just got feedback from a client that is clearly from ChatGPT" [E]. That reprices the agency's product from the buyer's side, not the seller's, and it arrives in a market where an agency principal warns "you can't win a price war with someone whose costs are near zero" [E]. The founder's defensive instinct is to double down on craft and human authorship, which is the wrong lever, because the client is not questioning who typed it [I]. The implication they have not drawn: the only thing left that a compliant model will never hand their client is the pushback [I].

**Consequence:** Scope-by-scope erosion as clients absorb the drafting layer and argue the retainer down to execution. Fee compression that shows up as slower renewals and smaller second-year contracts before it shows up as churn. And a relational shift in which the client arrives at the strategy call with a position already formed by an assistant that agreed with them, leaving the agency reacting instead of framing [I].
**Solution:** Hivemind's value here is not producing what the client's assistant produced faster. It is retrieval over a curated practitioner corpus that surfaces angles a general model trained on public writing will not reach, delivered through an interaction model that argues back and adapts when pushed [E]. That is the pushback the founder needs before the meeting, so they walk in with the counter-position rather than the same consensus answer the client already has. Two bounds to state plainly: the documented corpus is web3-scoped, with cross-industry applicability asserted only in an FAQ, and the product has no per-client memory, so the compounding asset is other practitioners' patterns, never a record of the founder's own [E].


# Pains And Frustrations

### 1) The best thinking you do all month goes out unpaid, and the paying account gets whatever's left

**Explanation:** The strategy work arrives with no line item attached, because clients "opt to skip strategic planning in favor of immediate implementation" and your own shop "does not reward non-billable activities" [E]. So it lands where it always lands: your evenings, your Sunday, the hours nobody is timing. You know what a proper job costs because you've priced it, "tens of thousands of dollars in internal resources and tons of time," which is exactly why you keep not doing it properly [E]. Worse, the deepest thinking in the building is systematically spent on prospects during pitch, unpaid and pre-signature, while signed accounts get the compressed version you assembled Thursday night. **(Hidden gem)** The inversion is invisible from inside because pitch strategy feels like selling, not like delivery, so it never shows up as capacity you spent [I]. Emotions: the low-grade resentment of doing your most valuable work for free, the private embarrassment of shipping thinking you know is thinner than your standard, and the fatigue of a week that never actually ends.

**Emotional Impact:** It costs you the weekend you promised someone, and it costs it repeatedly enough that you stopped promising. At 13% average after-tax margin you can feel every senior unbillable hour in the number you look at monthly (Promethean 2026, n=119) [E]. And when a client calls you their execution partner rather than their strategic one, you can't argue, because the record of what you actually charged for says they're right [I].

**Solution:** Hivemind is a monthly tool cost, not an engagement, covering positioning options, GTM sequencing, messaging frameworks and competitive differentiation in conversation [E]. That lets the routine account and the unpaid pitch get the same depth as the flagship, without a new invoice line you already know clients won't fund [I]. Qualify price on the call: no tier is published [E].

### 2) Every strategic decision in the shop is waiting on your next open block

**Explanation:** You can hand the work off. Agencies your size advertise "our AMs lead strategy," and yours do [E]. What you can't hand off is being the last check before it reaches a client, so the angle sits in a doc until you have ninety uninterrupted minutes, which this week you do not have. Hiring the peer who could catch it instead of you runs $150–500/hr or $4k–20k monthly, numbers that never pencil for the account paying you eleven grand a quarter (T2 marketplace pages, no stated n) [E]. So four briefs queue behind one calendar and your team learns to route everything upward, which makes the queue longer next month [I]. Emotions: the claustrophobia of being the ceiling on your own company, guilt toward a team stalled waiting on you, and the specific irritation of reading a good angle at 11pm and having no one to argue it with.

**Emotional Impact:** It shows up as the meeting you push for the third time and the Slack thread you leave unanswered because a real answer takes an hour you don't have. It follows you into the evening as the mental list of things only you can unblock. Over time it corrodes the reason you started the agency: doing the thinking, not being the constraint on it [I].

**Solution:** No onboarding step and conversation latency means the pressure-test happens in the hour the angle is written, not at the next open block [E]. Team and agency plans with shared team context let an AM run the same challenge before it reaches you, so you're checking a tested position rather than a first draft [E]. This buys capacity, not judgment [I].

### 3) You can smell it from the first line, and you refuse to be the person who sent it

**Explanation:** You use ChatGPT every day and you'd still say out loud, "I can spot AI-generated copy from a mile away. And I hate it… It's all starting to sound the exact same" [E]. The complaint is texture, not accuracy: "weak and fluffy," or in the harsher version circulating among your peers, "aggregated slop" [E]. So an AI strategy tool isn't a tooling decision for you, it's a question about what your agency is, and you'd rather stay slow than become interchangeable. The vendors don't help. The category's own pitch is escape from "the agency markup" and "$300/hr" advice, aimed squarely at the person paying you [E]. You have already drawn your own line, though, and it's a useful one: AI is fine "mostly for strategy or operations, never for creative" [E]. Emotions: contempt for the texture of the output, defensiveness about being lumped in with shops that ship it, and a quiet fear that the line you drew moves next year.

**Emotional Impact:** It costs you the relief you're entitled to, because the fastest available fix is the one you've ruled out on principle. It shows up as a slightly too-loud "we don't use AI for that" in a client meeting. And it leaves you carrying the bottleneck from Pain 2 for reasons that feel like integrity and read, from outside, like being slow [I].

**Solution:** Sell into the lane you already sanctioned. The claimed mechanic is adversarial, not generative: asked for a launch plan, it answers "Before we talk launch, what's your retention loop?… Walk me through the first 72 hours" [E]. Nothing it produces goes to the client. It pressure-tests the position you then write yourself [I]. Vendor claim, zero third-party verification [E].
### 4) You've now demoed four of these and they all said the same three things

**Explanation:** Every one of them told you it was trained on real practitioners rather than generic AI, and M1-Project says it in a customer's voice, and brand.ai promises briefs that aren't "generic frameworks" [E]. Every one of them told you its context compounds: Averi's "context never dies," Octave's context engine, Jasper IQ's context layer, AirOps' Brand Kit [E]. You cannot separate them on their own copy, so you fall back on the tab already open, which is what a paying reviewer of an adjacent tool concluded when he found no "significant advantage over free alternatives like GPT or Gemini" [E]. The cost of that isn't money, it's the three hours of evaluation you'll never get back and won't spend again next quarter. Emotions: the flatness of reading your fourth identical homepage, suspicion that the whole category is the same wrapper, and mild self-reproach for the hours you gave it.

**Emotional Impact:** It hardens you. The next founder who emails you about an AI strategy tool gets less patience than the last one, and the one after that gets none. Meanwhile the free tool keeps being adequate enough that nothing forces the decision, so the problem stays exactly where it was [I].

**Solution:** Don't argue corpus provenance or memory. Both are parity claims and Hivemind cannot make the second one at all [E]. Argue the one behavior no competitor's published positioning claims: a tool built to disagree with you rather than produce for you. It is demonstrable in the first conversation with no setup, which makes it a live proof rather than another homepage sentence [E].

### 5) Your client walked into the call already holding an answer, and it wasn't yours

**Explanation:** You braced for the wrong thing. The threat you rehearsed was your work sounding machine-made; the thing that actually happened was "I just got feedback from a client that is clearly from ChatGPT" [E]. Now the drafting layer you used to charge for is something they can approximate on their own, and you're negotiating scope against an operator whose marginal cost is nothing, in a market where a peer already warned "you can't win a price war with someone whose costs are near zero" [E]. Your instinct is to defend craft and human authorship, but nobody asked who typed it. What they're testing is whether your judgment is worth more than the answer their assistant already gave them, and that assistant agreed with them, which yours must not [I]. Emotions: the exposure of being reacted to instead of leading the room, indignation at reviewing feedback a model wrote, and the dread that this is the shape of every renewal from here.

**Emotional Impact:** It shows up as second-year contracts smaller than first-year ones and renewals that take three conversations instead of one. It changes how you enter a client call, from framing to defending. And it puts a question under your business plan that you don't say out loud, which is what exactly you are selling in eighteen months [I].

**Solution:** Retrieval over a curated practitioner corpus surfaces angles a general model trained on public writing does not reach, delivered through an interaction that argues back and adapts when you push [E]. You arrive with the counter-position rather than the consensus one already sitting in their inbox. Two bounds, stated plainly: the documented corpus is web3-scoped, with cross-industry reach asserted only in an FAQ, and there is no per-client memory, so what compounds is other practitioners' patterns, never a record of yours [E].


# Triggers

### 1) A client returns feedback on the positioning deck that was obviously written by ChatGPT

**Situation:** The framework went over Tuesday, three weeks of thinking compressed into eleven slides. Thursday morning the reply lands: a tidy, numbered, faintly ecstatic list of "considerations" with parallel structure on every bullet and not one specific reference to their own product. The founder recognizes the shape instantly, because they use the same tool every day and can spot it "from a mile away" [E]. One practitioner described exactly this moment in a peer forum: "I just got feedback from a client that is clearly from ChatGPT" [E]. What follows is not anger, it is a cold recalculation about what the retainer is actually paying for, because the client has quietly stopped treating strategic judgment as scarce. The fear underneath is commercial and immediate: if the client's assistant can review the work, the client will eventually ask why it can't originate it, in a market where "you can't win a price war with someone whose costs are near zero" [E]. They forward it to their business partner with no comment, then spend the evening trying to draft a counter-position sharp enough that no compliant model could have produced it [I].

### 2) They lose a pitch to a shop that walked in with a point of view instead of a discovery agenda

**Situation:** Two finalists, same room, same week. The agency's deck opened with process, team, and a discovery phase, because the real strategic work was quoted as phase one and could not be done for free. The other shop opened with an argument about the prospect's category and a positioning angle already drafted, and the prospect's team leaned in. Agency-consulting sources describe exactly this asymmetry: arrive with insight and you are the strategic consultant, arrive with a discovery plan and you "stumbled in with questions" [E] (source is vendor-adjacent, so treat the framing as a rebuttal to a real objection rather than neutral observation). The debrief email is polite and says the other team "felt further ahead on thinking," which is the sentence the founder replays on the drive home. It confirms the private worry a consultant put bluntly about a founder in this exact spot: an inability to explain why his agency was different from cheaper alternatives [E]. Within a week they start asking peers how anyone produces a defensible point of view speculatively, across four live pitches, without burning a senior week per opportunity [I].

### 3) A launch date moves up and the messaging framework was never actually written

**Situation:** The client's product ship slips forward six weeks to hit a board meeting, and the agency's content calendar, paid brief, and sales one-pager all inherit the new date. Every one of them is downstream of a positioning decision that exists as three bullets in a kickoff doc and a conversation nobody transcribed. The team starts producing anyway, because the retainer is execution and execution is what gets invoiced, while the strategic layer stays unfunded and therefore unscheduled [E]. By Wednesday the founder is reading two email drafts that describe the product as two different things, and the felt experience is not panic but a specific dread of the review call where someone asks "so what's the angle?" and there is no answer that survives a follow-up question. Doing it properly is not the option on the table, because an agency principal put the cost of custom brand strategy at "tens of thousands of dollars in internal resources and tons of time" [E]. So the founder blocks Saturday, writes the framework alone, and afterward goes looking for something that could have argued with them at 9pm Wednesday instead [I].

### 4) A renewal comes back with the strategy line struck out and the execution hours intact

**Situation:** The scope arrives redlined by someone the founder has never met, and the pattern is unmistakable: retainer hours preserved, the strategy and planning allocation removed, and a note about doing that piece internally this year. It is not a churn event, which is what makes it harder to react to, and it lands on an average after-tax margin of 13% where the removed line was the highest-value hour in the shop [E]. The client is doing what the demand side generally does, which practitioners describe as opting "to skip strategic planning in favor of immediate implementation" [E]. What the founder feels reading it is a kind of demotion, the quiet reclassification from partner to vendor, and the fear that this is the template the account team will now propose to every other client at renewal. They know the counterargument, that clients pay for "expertise, strategic judgement and business outcomes" rather than hours [E], but they have no cheap way to demonstrate it between now and the countersignature. They start pricing what it would cost to bring a defensible strategic point of view to every renewal conversation next quarter [I].

### 5) A client's legal team sends over an AI clause and the founder realizes what has already been pasted into ChatGPT (hidden gem)

**Situation:** It arrives as an addendum, not a negotiation: disclosure of any generative AI used in deliverables, restrictions on where client material may be processed, and a representation the founder is asked to sign. Practitioners report exactly this shift, with clients' own AI policies now including clauses agencies are asked to agree to, applied to decks that "contain proprietary client info" [E]. The founder signs nothing yet, because the honest answer requires asking six people what they have been doing, and the honest answer is that decks have been going into ChatGPT for months to find information buried across hundreds of slides and to rewrite headlines [E]. The exposure is retroactive, which is the part nobody prices: the clause governs future conduct, but the questionnaire that accompanies it asks what has been used to date. The feeling is not guilt so much as the specific vertigo of discovering your own answer is unknowable, and the fear is a single blocking email from a client's counsel that stalls an entire account. Note what this does to vendor selection: the next tool the founder evaluates gets asked about data handling before it gets asked about output quality, and Hivemind currently publishes no SOC 2, DPA, or training-opt-out statement, which makes this a trigger the seller must be prepared to lose on [E].

### 6) The senior who ran strategy on three accounts gives notice

**Situation:** She was the reason the founder could stay off two of the four weekly strategy calls, and the posting that hired her promised exactly that, the way one live agency listing advertises "our AMs lead strategy" [E]. The resignation is amicable and the notice period is four weeks, which is the problem, because three clients have quarterly planning inside that window and the replacement search for someone with genuine positioning judgment realistically runs a quarter. The alternative is the market rate the founder already knows, roughly $150 to $500 an hour for fractional senior help and $4,000 to $20,000 a month on retainer, economics that work for the flagship account and nothing else [E]. What the founder feels in the days after is the return of a familiar constriction, the sense of every strategic question routing back to one calendar. The fear is quality drift: work going to clients that they never pressure-tested and would have caught. They start looking for something that lets a remaining AM stress-test their own thinking before it reaches the founder's inbox [I].

### 7) A junior turns around a strong strategic deliverable in two days, and the founder finds out how (hidden gem)

**Situation:** The work is genuinely good, which is why the founder asks about it. The answer is that the strategist built a custom assistant loaded with that client's transcripts, past decks, and brand guidelines, and has been running her thinking through it for a month. At least one agency has productized this pattern, giving each client a custom GPT that understands their specific context [E] (a single documented case, not a segment norm). The founder's first reaction is relief and the second, arriving about an hour later, is a proper chill: the shop's strategic quality bar has been delegated to an unlogged, unreviewed, personally-configured model that no one else can inspect, which is precisely the failure a growth publication names when it warns "you can't protect your brand voice if everyone's prompting in different tools" [E]. The political dimension is what makes it hard to act on, because banning it punishes the best output the team produced this month and admits the founder was not the one who found the leverage. The fear is not that the tool is bad, it is that the founder no longer knows which client deliverables were argued with and which were merely agreed with. They start looking for a sanctioned option the whole team can use with shared team context, so the pressure-testing is visible rather than private [I].

### 8) Six months after paying a brand strategist, the founder is asked what it bought (hidden gem)

**Situation:** The engagement produced everything it promised: exercises, ideal client profiles, value propositions, a deck, a tagline. A consultant describes the aftermath for a founder in exactly this position, six months on with the same clients and the same pricing pressure, still carrying the same inability to explain why his agency was different from cheaper alternatives [E] (the account is published by a rival positioning firm and is marketing, not neutral evidence). The moment it becomes intolerable is not the discovery, it is the partner meeting where someone asks, without malice, what changed. The founder is defending a five-figure decision they made, championed internally, and cannot point to a single won pitch that traces back to it, which is the specific humiliation of having bought the credible thing and gotten nothing. This is why the next purchase in this category gets evaluated on whether it changes what happens in a live pitch next month rather than on the quality of the artifact it produces. Note the seller's exposure at this exact moment: Hivemind has no case study, named customer, review-site presence, or quantified outcome anywhere, so this founder will be asking for proof the vendor cannot currently supply [E].

### 9) The founder pushes their own strategy block for the third week running while headcount is at an all-time high

**Situation:** The recurring Wednesday afternoon block exists for thinking about positioning, on client accounts and on the agency's own. It has been moved for a delivery escalation, a pitch rehearsal, and a hiring conversation, in that order, during a quarter where the team is the largest it has ever been. The founder notices the pattern only when the calendar invite turns into a fourth reschedule, and what surfaces is not exhaustion but something closer to alarm at the shape of the business, which advisors to this audience frame as a firm running on the owner's memory and constant presence rather than on systems [E] (advisor content, not buyer testimony). The intolerable detail is that headcount went up and strategic throughput did not, meaning the constraint was never staffing. It was that there is nobody to argue with at the moment the thinking happens, and the internal comp structure "does not reward non-billable activities," so no one else is incentivized to be that person either [E]. They message two peers in an agency-owner Slack to ask what anyone is actually using for this, and read the replies at 11pm [I].


# Barriers

### 1) "I already pay for ChatGPT. What does this do that it doesn't?"

**Situation:** This is the first question, not the last, and it arrives before any demo. The founder has a paid assistant open on a second monitor and has already used it that morning for something strategic. Every paid alternative they have tried made the same argument and lost to the incumbent. [I]

**How To Spot:** The proxy-category verbatim is exact: a paying reviewer of an adjacent AI tool concluded the features "didn't provide any significant advantage over free alternatives like GPT or Gemini." [E] Watch for the founder asking for a feature list rather than a working session, asking "what's it trained on" as a substitute for evaluation, or saying they'll "play with it this weekend." These signals route the decision to a solo comparison against a tool that costs them nothing incremental. [I] Note the literal phrasing "how is this different from ChatGPT" is unattested in public buyer text; do not put those words in their mouth on a call. [E]

**Specific Stopper:** The founder never runs a comparison at all. They price the delta against a subscription they already pay for, find no observable difference in a five-minute prompt test, and close the tab.

**How To Handle:** Refuse the feature comparison. Run the demo behavior live, in their account, on their live client problem: ask them for the launch plan they need, and let the tool decline it and ask the prior question. The published pattern is "Before we talk launch, what's your retention loop?… Walk me through the first 72 hours." [E] Then have them paste the identical prompt into ChatGPT in the same session. The difference to sell is refusal, not knowledge. Do not claim corpus provenance as the differentiator; M1-Project runs that exact contrast in customer voice already. [E]

### 2) "AI strategy is fluffy. My clients pay me not to hand them that."

**Situation:** Output-quality skepticism here is aesthetic and identity-shaped, not accuracy-shaped. This founder can identify machine text on sight and finds it contemptible. It surfaces the moment they read a full response rather than a headline. [I]

**How To Spot:** In their own words: "I can spot AI-generated copy from a mile away. And I hate it… It's all starting to sound the exact same." [E] Practitioners describe the specific defect as "The writing is very weak and fluffy" and, more bluntly, "a computer that just spits out aggregated slop." [E] Behaviorally: they skim the first two paragraphs of an output and stop; they ask whether it can match their voice; they forward one output to a colleague with no comment. The permission structure that cuts your way is also verbatim: "Yes but mostly for strategy or operations, never for creative." [E]

**Specific Stopper:** One generic-sounding paragraph in the first session confirms the prior, and the tool is filed as a content generator with a strategy label.

**How To Handle:** Sell into the sanctioned lane and stay there. Frame every session as back-of-house pressure-testing before a pitch, never as drafting client-facing work. You are matching the line they already drew. Structure the trial around three questions with a defensible right answer they already know (a positioning call they got right last year), so they grade judgment, not prose. Do not demo the landing-page or email-sequence capability in the first session even though it exists; it invites exactly the grading criterion that loses. [I]

### 3) "Every one of these tools just tells you what you want to hear."

**Situation:** The second objection under the first. Even founders who accept the tool isn't slop assume it is agreeable, which makes it worthless as a second opinion. It surfaces immediately after you claim the tool argues back. [I]

**How To Spot:** The practitioner phrasing is specific: "Its sycophant mode has been a real problem of late." [E] They test it by asserting something deliberately weak and seeing whether it agrees. They ask "does it ever tell me I'm wrong?" They describe their custom GPT as useless because it "just reproduces my own brief back at me," the documented workaround failure. [I]

**Specific Stopper:** They feed it a bad angle, the model validates it, and the entire adversarial positioning collapses in one exchange with no recovery.

**How To Handle:** Run that test for them rather than waiting for it. Bring a deliberately flawed positioning statement from a real B2B software company and ask the tool to validate it on the call. Then hand them the keyboard and tell them to try to get it to agree with something they know is wrong. The published mechanic supports this: it is built to challenge assumptions and to adapt when pushed back on. [E] If it caves, you have lost that deal honestly; do not manage around it, because this buyer will run the test in week one regardless. State the limit out loud: this is a vendor claim with no third-party verification anywhere. [E]
### 4) "Who else is using it? Show me an agency like mine."

**Situation:** Standard mid-cycle due diligence that this product cannot survive on paper. There are no named customers, no case studies, no G2 or Capterra listing, no quantified outcome, and no attributed testimonial at any source tier. [E] It surfaces the moment the founder starts thinking about a team rollout rather than a personal trial.

**How To Spot:** "Send me a case study." "Who's your reference in the B2B SaaS agency world?" They Google the product and hit a name collision with a recruiting tool and a Notion template. [E] They ask for a customer to talk to and go quiet when none arrives. Watch for the diligence pattern that predicts a stall: a request for logos followed by a two-week silence. [I]

**Specific Stopper:** A founder who cannot name one peer using it will not put it in front of their own team, because the internal cost of championing an unproven tool exceeds the tool's price.

**How To Handle:** Stop performing proof you don't have. Say plainly that there is no case study yet and turn the gap into terms: offer a design-partner arrangement with a named end date, direct access to the Myosin practitioners behind the corpus, and a written commitment that their outcome becomes the first published case study only with approval. The parent agency's engagements (Stellar, Provenance Blockchain) are human-delivered agency work with zero attribution to this product, so never present them as product proof. [E] The one asset you do have is the first session itself; make it the reference.

### 5) "I can't approve a number you won't give me."

**Situation:** No price is published at any tier. The only public packaging fact is that individual, team and agency plans exist. [E] This surfaces at the exact point the founder is warmest, right after a good first session, which is why it costs more than it looks.

**How To Spot:** "What's this going to run me per seat?" "Is there a trial?" "What happens if I put three people on it?" They ask for pricing in writing to compare against a tool they already budget for. They stall on a rollout decision they had verbally agreed to. [I] The comparison set is unforgiving: adjacent tools publish entry points in the $45–$99/month range. [E]

**Specific Stopper:** With no number, the founder cannot slot the tool into the mental category "routine monthly software," so it gets evaluated against the fractional-strategist band instead, $150–500/hr, $4k–20k monthly retainers, and loses on risk. [E]

**How To Handle:** Quote a firm monthly number on the first call and put it in writing the same day. Anchor it explicitly as tool budget, not strategy budget: name the per-account monthly cost across their client book. Avoid the seller's own published anchor: "stop paying $300/hr… without the agency markup" argues against the buyer's business model and invites them to price you against their own rate card. Offer a one-month paid pilot with a stated cancel path rather than a free trial; free trials in this band die of non-use, not of price.

### 6) "That's client-confidential material and I'd be pasting it into a vendor I can't vouch for."

**Situation:** The material the founder would bring is client discovery, positioning work, and competitive context they are contractually holding. Practitioners report that decks contain proprietary client information and that clients' own AI policies now carry clauses agencies are asked to agree to. [E] It surfaces the second the founder considers using it on a real account rather than a hypothetical.

**How To Spot:** "Where does this go? Do you train on it?" "My client's MSA has an AI clause." "Can you send whatever security doc you have?" [I] Behavior: they run the trial on a fictional company or a former client, never a live one, which caps perceived value and guarantees the pilot underperforms. There is no SOC 2, DPA, data-residency statement, or training-opt-out published. [E]

**Specific Stopper:** A single unanswered email about training data converts an active evaluation into permanent silence. This band has no procurement process to escalate into, so the objection has nowhere to go but dead.

**How To Handle:** Pre-empt it in the first session before they ask. Have a one-page written answer ready covering retention, training use, and deletion on request, signed by a named person at Myosin, and send it unprompted. Where the answer is "we don't have that yet," say so and offer contractual language instead of a certification. Separately, give them a sanctioned low-exposure path: run the first three sessions on their *own* agency's positioning, which is confidential to nobody but them and is also the pitch problem they most need solved. [I]
### 7) "Everything it gives me stays in the chat window." (hidden gem)

**Situation:** The founder's measure of a strategy artifact is whether it can be shown, billed against, or pasted into a pitch. This product has no export, no document output, no templates, and no integrations of any kind: not Notion, not Google Docs, not Slack. [E] The barrier surfaces in week two of use, not during the demo, which is why it reads as churn rather than objection.

**How To Spot:** "Can I get this into Notion?" This is not a nice-to-have for this segment: an agency of exactly this profile states "Notion is the center of our content operation… Every article starts in Notion." [E] Then: "How do I share this with my AM?" Behaviorally, they copy-paste one output into a doc, don't come back, and answer your check-in email with "haven't had a chance." Compounding it, the tool holds no per-client memory, so returning after ten days means re-briefing the account from zero. [E] The Exit Five articulation of the failure mode is in-house voice, not agency voice, but the shape is the same: the work gets done and "then it just sits in a Google Doc." [E]

**Specific Stopper:** The re-briefing tax quietly exceeds the value of the session, so the founder's usage decays to zero without a cancellation conversation. The deal is lost to non-adoption, never to a competitor.

**How To Handle:** Design the pilot around this constraint instead of hoping it doesn't bite. Give the founder a saved brief block, one paragraph per client, pasted at the top of every session, and build it *with* them for their three largest accounts on the onboarding call, so the re-brief costs fifteen seconds instead of ten minutes. Set a standing weekly slot tied to a real recurring event (Monday pitch prep, Thursday client strategy call) rather than telling them to "use it when stuck." Never imply per-client memory exists; sell the paste-block as the workaround it is. [I]

### 8) "This thing was built for token launches, not for my SaaS clients." (hidden gem)

**Situation:** Any founder who does thirty seconds of diligence lands on public surfaces describing a curated Web3 marketing knowledge base, a DAO, and nailing positioning "before the token goes live." [E] Trigger conditions in the documented skill spec include token launch tactics and community growth. [E] This surfaces after the good first call, when they research alone.

**How To Spot:** "Is this a crypto thing?" "Do you have anything in B2B SaaS?" [I] Behavior: they go quiet after visiting the site, or they forward a screenshot of the homepage with a question mark. Cross-industry applicability is asserted publicly only inside an FAQ answering whether it's crypto-only, an answer to an objection, not a positioning. [E]

**Specific Stopper:** The founder concludes the practitioner corpus contains nobody who has run their kind of account, which kills the single claim the product rests on: that the advice comes from people who have actually shipped this work.

**How To Handle:** Get ahead of the URL. Name the crypto heritage in the first two minutes and reframe it as adjacency, not mismatch: pre-product-market-fit launches, narrative-led categories, community-driven growth, and story-market-fit problems, which is the shape of most Series A–B software launches too. Then prove it in-session on their live account rather than arguing it. Ask them for their hardest current client positioning problem and let the output be the evidence. If the response returns crypto-shaped examples, do not defend it; log it and price the deal accordingly. [I]

### 9) "If my team sees a tool that promises strategy without the agency markup, what am I telling them?"

**Situation:** Internal role threat, and it is self-inflicted by the seller's own copy. The published anchors, "stop paying $300/hr," "without the agency markup," strategic firepower "that used to cost six figures," argue against the exact expertise the founder's senior people bill for. [E] This surfaces at rollout, after the founder personally likes the tool.

**How To Spot:** The founder buys a single seat and never expands. "Let me use it myself for a bit first." Senior strategists and AMs granted access never log in. Watch for the cultural tell already circulating in agency forums: "the only reason agencies say they are adopting AI is because other agencies are saying," and the unresolved internal question "is it ethical or just efficient?" [E] Team plans exist with shared context, so a stalled expansion is a signal, not a packaging gap. [E]

**Specific Stopper:** The team reads the tool as the business case for their own redundancy and declines to adopt it, so the account stays at one seat and renews on one person's habit, the most fragile renewal there is.

**How To Handle:** Rewrite the internal pitch before the founder makes it. Give them a short framing to send their team that positions the tool as the senior peer none of them has, used to pressure-test *before* it reaches the founder. This converts it from a threat to the AM into a route around the founder bottleneck. On the rollout call, seat the senior strategist first, not the founder, and have them run the deliberately-flawed-angle test from Barrier 3 so their first experience is being argued with, not replaced. Never repeat the "$300/hr" or "agency markup" language in front of anyone at the agency. [I]
# Alternative Ways

### 1) Strategy is the most senior work in the shop and the least billable, so it lands in founder hours or gets skipped (senior unbillable time eats a 13% average after-tax margin, and the agency reads as an executor at renewal)

**Alternative Ways:**
- Sell a paid diagnostic before any execution proposal: the Win Without Pitching / Blair Enns model where the engagement starts with a priced discovery, not a free scoping call [E].
- Productize the layer as a fixed-fee sprint with a named deliverable, the way Fletch PMM sells two-week positioning sprints to B2B software companies [E].
- Run April Dunford's *Obviously Awesome* ten-step process as a billable workshop, with competitive alternatives and value themes as the client-facing artifact [I].
- Hire or promote a Strategy Director who carries a utilization target, so strategic hours have an owner and a rate rather than defaulting to the principal [I].
- Rebuild the SOW so positioning and GTM sequencing are line items with hours attached instead of assumed inside a monthly retainer [I].

**Non Mainstream Solutions:**
- Tag every non-billable hour by client in Harvest or Toggl for one quarter, then rank accounts by true margin and reprice or fire the bottom two [I].
- Reverse-scope the renewal: present next year as strategy-first with execution as priced options, forcing the client to name what the layer is worth [I].
- Keep a per-account decision log: date, the call made, the alternatives rejected, the assumption it rests on, and read it back at renewal as evidence of what you were actually paid for [I].
- Harvest completed frameworks from finished accounts into an internal pattern library, so the second use of a positioning structure costs a third of the first [I].
- Score every lost pitch at 90 days on a single forced question: lost on strategic depth or lost on price. One owner, written, circulated to the team [I].

### 2) The founder is the strategist and the invoice-signer, so strategic capacity is bounded by one calendar (angles stall or ship unchallenged, and fractional help is priced out of every account but the flagship)

**Alternative Ways:**
- Bring in a fractional CMO through MarketerHire or GoFractional at $150–500/hr or a $4k–20k monthly retainer for the accounts that can carry it [E].
- Join a moderated peer group: Setup® Agency Mastermind runs forums of six to eight non-competing agencies scoped to 10–50 FTE shops, for structured second opinions [E].
- Install EOS/Traction's accountability chart and move the founder out of the delivery seat by naming a single accountable owner for strategy [I].
- Let account managers lead strategy with a founder review gate, which is already the advertised model at agencies this size [E].
- Retain a positioning consultancy on a per-engagement basis for the accounts where the angle is genuinely contested [I].

**Non Mainstream Solutions:**
- Red-team pass: a senior who did not build the angle argues the opposing case for thirty minutes before anything reaches the client, with the strongest objection written down [I].
- Forced-choice presentation: never one positioning, always two mutually exclusive ones with the tradeoff stated, which moves the founder from author to decider [I].
- Cap founder strategy review to two fixed 90-minute windows a week; work that misses the window ships on the AM's judgment, making the bottleneck measurable instead of ambient [I].
- Pre-mortem before ship (Klein's method): assume the angle failed in six months and have two people write the reason unprompted [I].
- Standing swap with one non-competing agency founder: an hour a month of mutual pressure-testing under NDA, no invoice either direction [I].

### 3) AI is the obvious relief valve, and this founder recognizes its output on sight (an entire tool category gets ruled out on identity grounds while the bottleneck stays intact)

**Alternative Ways:**
- Write an AI usage policy with an explicit human-authorship line, and put a disclosure clause in the MSA before a client asks for one [I].
- Build a brand-voice custom GPT per client loaded with the style guide, the approach documented at agencies trying to stop voice drift across tools [E].
- Confine generative tooling to the sanctioned lane this audience already accepts: strategy and operations, never front-of-house creative [E].
- Add a senior editorial QA pass as a required gate before anything reaches a client inbox [I].
- Route AI use to research synthesis across decks and transcripts rather than drafting, which is the use practitioners already describe doing openly [E].

**Non Mainstream Solutions:**
- Blind slop test: mix three AI drafts with three human drafts and have the team label them cold. If the hit rate is near chance, the objection is aesthetic, not commercial [I].
- Maintain a living tell-list of the shop's own AI fingerprints (em dash cadence, tricolons, "it's not X, it's Y") and enforce it at the edit stage, not the brief stage [I].
- Make the client's own AI policy a standard onboarding question, logged per account, so the constraint is known before scope is priced [I].
- Write a craft carve-out into the SOW: named deliverables produced without generative tooling, converting a private anxiety into a contract term the client can see [I].
- Publish the shop's AI stance as positioning and run it in two live pitches against a control pitch that stays silent on it [I].
### 4) Every alternative makes the same three claims in the same words (comparison collapses back to the free tool already open on the desktop, and the trial never converts)

**Alternative Ways:**
- Run a structured trial with success criteria written before signup rather than after the demo [I].
- Compare within the review categories that exist, G2 added AI Marketing Agents in May 2026, and read the one-star reviews first [E].
- Ask the peer Slacks (Grow Your Agency, Exit Five) for named recommendations from operators with the same client mix [E].
- Bake off free tiers against each other on the same brief before any card goes in [I].
- Name a single tool owner and a monthly spend threshold under which nothing gets bought without approval, the norm advised for agencies this size [E].

**Non Mainstream Solutions:**
- Same-brief bake-off: one real live client brief run identically through every contender and through the free incumbent, outputs anonymized and scored blind by two seniors [I].
- Score on disagreement rate rather than output quality. Count how often the tool contradicts the brief's premise; a zero contradicts nothing and buys nothing [I].
- Strip branding from four vendor homepages and have the team attribute them; anything unattributable is table stakes and scores zero on differentiation, the method 3C Ventures used on agency platform pitches [E].
- Buy-two-kill-one: adoption is conditional on naming the tool it replaces at signup and cancelling that one inside 30 days [I].
- Set a 30-day activation checkpoint instead of a renewal date. If fewer than three named people touched it on a real account, cancel, because tools in this price band die of non-adoption rather than competitive loss [E].

### 5) The client now has AI too, and judgment is being repriced from the buy side (scope erodes toward execution, and renewals shrink before they churn)

**Alternative Ways:**
- Move up-market to buyers whose internal teams have no capacity to absorb the drafting layer [I].
- Niche harder into a vertical so the pattern knowledge is the product and the drafting is incidental [I].
- Shift to value-based or outcome-linked pricing so the fee stops tracking hours the client can now replicate [I].
- Add a measurement or analytics service line, which trade coverage frames as the "business outcomes" clients say they are actually paying for [E].
- Run QBRs with attribution reporting so the retainer conversation opens on results rather than deliverable counts [I].

**Non Mainstream Solutions:**
- Bring the client's own AI into the room: run their prompt live on the call, then walk through what it missed and why the corpus it drew on could not know [I].
- Ship one unrequested contrarian memo per account per quarter, arguing against the client's stated plan on a single page [I].
- Log "client sent us AI-written feedback" as a formal account event and treat it as a renewal-risk flag with a 60-day intervention, since the inversion is already happening in practice [E].
- Build rate-card asymmetry: price execution at market and the diagnostic above it, so the compressible layer is visibly the cheap one [I].
- Run win/loss interviews at renewal through a third party, never the account lead, with one required question: what do you now do in-house that we used to do [I].


# Already Know

### Known Solutions

1. **Positioning frameworks.** The canon this founder has already read: April Dunford's *Obviously Awesome* and *Sales Pitch*, Geoffrey Moore's value proposition template, Ries and Trout, and Fletch PMM's homepage and messaging patterns distributed as free Figma and Notion templates [E]. These are diagnostic scaffolds, not decision engines, they tell you which boxes to fill, never which answer is right [I].

2. **Messaging artifacts.** Message house, message map, messaging matrix, messaging framework, positioning guide, the segment uses at least five names for overlapping documents, and Cascade Insights has published explicitly to disambiguate them [E]. Product Marketing Alliance and Exit Five both distribute member templates, and practitioners advise picking one and staying consistent so stakeholders argue about "the meat instead of template" [E].

3. **GTM planning documents.** ICP definitions, persona docs, launch tiering, channel sequencing, and the GTM plan itself, usually assembled in Google Slides or Notion from MKT1 and PMA template libraries [E]. Ignition sells the purpose-built version of this workflow to product marketers, research through launch planning through measurement [E].

4. **General-purpose AI assistants.** ChatGPT and Claude are installed in every shop in this segment and appear as a listed skill on live agency job postings alongside Google Ads and HubSpot [E]. They are used for deck retrieval, headline rewrites, and cross-research theme-finding, synthesis over material the founder already supplied [E].

5. **AI content platforms.** Jasper, Copy.ai, AirOps, and Surfer occupy the volume layer, with Jasper positioning around brand-consistent output and Copy.ai around consolidating "dozens of copilots and point solutions" into one GTM platform [E]. Every one of them is an execution tool wearing strategy language; none is built to refuse a brief [I].

6. **AI context and brand-governance layers.** Octave's "GTM Context Engine," Jasper IQ's "context layer for marketing content," AirOps Brand Kit as "the governance foundation for the AI era," brand.ai, and Averi's workspace "where context never dies" [E]. All four assume one brand per workspace, which is why none of them reads as agency-native to a founder running eleven client brands [I].

7. **Agency operating systems.** Notion as the system of record, one agency states plainly that "every article starts in Notion," plus Asana, ClickUp, Workamajig, Function Point, and Teamwork for delivery, with ZenPilot selling the process layer on top [E]. These make execution legible and do nothing for the quality of the thinking that precedes it [I].

8. **Buying human judgment.** Fractional CMOs at $150–500/hr or $4k–20k monthly retainers, senior freelance strategists, and packaged positioning engagements such as Fletch PMM's two-week sprints, which claim 500+ B2B software companies (T2 marketplace and vendor pages, no stated n) [E]. Adjacent to these sits Wynter, which sells external validation rather than the strategy itself: verified B2B buyer reaction in under 48 hours, credits at $1 each [E].

9. **What the top decile actually runs.** A custom GPT per client, loaded with that client's context and shared with the client, the documented case is Develomark, whose workaround escalated into building an internal tool called SplashDash to unify access **(Hidden gem: the sophisticated workaround does not stay a workaround; it acquires an owner, a backlog, and a maintenance line, quietly converting a 25-person agency into a small software company nobody staffed for)** [E]. Alongside it: CLAUDE.md-style context files and MCP servers wired into Claude Code, Clay for job-change and funding triggers, Exit Five's guidance to feed positioning docs into your own AI as "training that never stops," and productized AI-transformation programs sold to agencies at $15,000/month on three-month minimums [E].
### Solution Attempts

#### 1) Built a canonical positioning template in Notion and mandated its use across accounts
**Setbacks:** A template standardizes the shape of the output without raising the quality of the thinking, so the founder still ends up filling the hardest boxes personally at 9pm before a pitch [I].
**New Problems:** Every client now has a half-completed strategy doc in the workspace, which reads to the team as evidence the work was done and quietly removes the pressure to finish it [I].

#### 2) Dumped client decks and research into ChatGPT to accelerate synthesis
**Setbacks:** It returns what was already in the deck, restructured. Practitioners describe the output as "weak and fluffy" and the model's agreeableness as "a real problem of late" [E].
**New Problems:** Proprietary client material is now sitting in a general assistant with no answer prepared for the client-side AI clause that lands in the next MSA [E].

#### 3) Built a custom GPT per client, loaded with that client's context
**Setbacks:** A model fed exclusively the founder's own material can only reflect the founder's own position back. It is retrieval, not a second opinion [I].
**New Problems:** Eleven GPTs now need refreshing every time a client repositions, and the maintenance sits with whoever built them, usually the person the shop can least afford to have doing it [I].

#### 4) Hired a fractional CMO or senior freelance strategist on retainer
**Setbacks:** At $150–500/hr the economics only clear on the flagship account, leaving the mid-tier book exactly as unresourced as before [E].
**New Problems:** The retainer becomes a fixed monthly cost that must be justified to a 13% margin, so it gets allocated to billable client work and stops being available for speculative pitch thinking [I].

#### 5) Bought a two-week positioning sprint from a named consultancy
**Setbacks:** The output is one deck for one client, delivered once, with no mechanism for the next eight accounts [I].
**New Problems:** The founder has now demonstrated to the client that the real strategic authority sits outside the agency, which is a difficult impression to reverse at renewal [I].

#### 6) Bought an AI content platform expecting the strategy layer to come with it
**Setbacks:** The platform optimizes for on-brand volume, and a paying reviewer of exactly this category concluded the features "didn't provide any significant advantage over free alternatives like GPT or Gemini" [E].
**New Problems:** A per-seat subscription now sits on the P&L producing drafts nobody was short of, which poisons the internal case for the next tool the founder actually needs [I].

#### 7) Delegated strategy to senior account managers so the founder could step back
**Setbacks:** The delegation is real, one live agency posting advertises "our AMs lead strategy," but the founder remains the quality bar, so every angle still routes upward for approval [E].
**New Problems:** The bottleneck relocates to a review queue instead of disappearing, and the AM learns that their strategic work exists to be corrected rather than shipped [I].

#### 8) Ran a discovery workshop to build consensus before writing anything
**Setbacks:** Workshops manufacture alignment, not insight, and practitioners describe the socialization path as strictly serial: one-to-few, then team, then revenue org, then C-level [E].
**New Problems:** The calendar cost is unbillable and lands on the founder's week, converting the exact hours the exercise was meant to protect into more of the unfunded strategic layer [I].

#### 9) Wrote an AI usage policy and banned it from client-facing craft
**Setbacks:** Governance stops the visible risk and does nothing about the actual constraint, which is access to senior judgment, not access to generation [I].
**New Problems:** Usage moves off the sanctioned path into individual personal accounts, so the founder loses the visibility the policy was written to create, the pain a practitioner names as not being able to protect brand voice "if everyone's prompting in different tools" [E].

#### 10) Packaged strategy as a paid discovery sprint to make it billable
**Setbacks:** Clients "opt to skip strategic planning in favor of immediate implementation," so the sprint gets negotiated out of the SOW or discounted to a token line item [E].
**New Problems:** Once strategy carries a price the client declined, the agency has publicly conceded it is optional, which weakens every future attempt to charge for it [I].

### Non Mainstream Ideas

**"The only AI in your stack that tells you no."** Every competitor sells compliance dressed as capability: faster ICPs, more on-brand copy, context that never dies. Hivemind's demo does the opposite. Asked for a full launch plan, it refuses and asks "Before we talk launch, what's your retention loop?" [E]. Refusal is demonstrable in a first conversation with no onboarding, which makes it a live sales asset rather than a copy claim, and it is the one behavior on the parity list nobody else asserts [I].

**"Back of house. Never front of house."** This audience has already written its own permission slip: AI is acceptable "mostly for strategy or operations, never for creative" [E]. Position explicitly inside that boundary rather than around it. The product is what the founder consults before writing, never what the client reads. Naming the limit is more credible than claiming there isn't one, and it neutralizes the slop risk without arguing with it [I].

**"Priced for the eight accounts that never got a strategist." (Hidden gem)** The flagship client already gets senior thinking; the mid-book gets whatever the founder can reach by Thursday. The real conversion risk in this ACV band is not competitive loss but non-activation: deals under $25k are documented as dominated by activation rather than evaluation (T3, treat as hypothesis) [E]. So sell against a *specific named account* the founder is currently under-serving, not against a workflow, because a tool tied to one at-risk logo gets opened in week two [I].

**"Your client already asked ChatGPT. Come with the counter-position."** The founder's rehearsed fear is producing machine-sounding work; the attested event is receiving client feedback "clearly from ChatGPT" [E]. Reframe the purchase as pre-meeting preparation: the client arrives with a consensus answer, and the agency's job is to have already argued with it. Retrieval over a practitioner corpus is defensible here in a way corpus-provenance marketing is not, because the claim is about reaching a different angle, not a better sentence [I].

**"We publish the arguments, not the case studies."** There are no named customers, reviews, or quantified outcomes at any source tier, and manufacturing them takes quarters [E]. Invert the gap: publish real transcripts where the product pushed back on a live positioning question and the operator disagreed. This is the only proof format that demonstrates the differentiating behavior directly, and it is structurally unavailable to competitors whose product is agreeable output [I].

**"Buy the argument, not the archive."** Persistent memory is table stakes and Hivemind cannot claim it: no per-client memory, no workspace, no export [E]. Turn the constraint into the pitch: the founder does not need a system that remembers what they already think, which is what a custom GPT loaded with their own decks already does and what six vendors are selling. They need access to how other practitioners solved it, which is a corpus question, not a memory question [I]. State the web3 scope of that corpus plainly on the call rather than letting a buyer discover it in week three [E].
# Solution Selection Criterion

### 1) Adversarial challenge: impacts whether the founder walks into a client meeting with a position they've actually stress-tested

**Importance:** 9/10. The vendor's own demo and the buyer's contempt for agreeable AI are both verbatim-supported. [E]
**Explanation:** This buyer already has infinite generation capacity and no senior peer to tell them the angle is wrong; practitioners in this audience name sycophancy as the failure ("Its sycophant mode has been a real problem of late"). [E] A tool that executes the request is worthless to someone whose scarcity is judgment, not output. [I]
**Purchase Decision:** They ask for a full launch plan on a live account and watch whether the tool delivers it or refuses and asks the prior question. The published demo answers "Before we talk launch, what's your retention loop?… Walk me through the first 72 hours." [E] Then they push back on a recommendation to see if it caves. Internally: *"Did it tell me something I didn't want to hear, or did it just flatter my brief back at me?"*

### 2) Output register: impacts whether any of this can survive contact with a client

**Importance:** 9/10. The fear is stated first-person and with a named tell. [E]
**Explanation:** The buyer's objection is not accuracy, it is texture: "I can spot AI-generated copy from a mile away. And I hate it… It's all starting to sound the exact same," and separately, "weak and fluffy" and "aggregated slop." [E] Being caught producing slop reprices the one thing they bill for. [I]
**Purchase Decision:** They paste a paragraph of the output into a doc and read it aloud, hunting for consensus phrasing and hedge stacks; several will run it past a senior writer without saying where it came from. Internally: *"If this landed in my inbox from a junior, would I know?"*

### 3) Corpus relevance to B2B software: impacts whether the recommendations apply to their actual client book

**Importance:** 9/10. The documented knowledge base is explicitly scoped to web3, and cross-industry fit is asserted only in an FAQ. [E]
**Explanation:** The differentiating claim is a curated practitioner corpus, and the only documented corpus is "a curated Web3 marketing knowledge base" with triggers like "token launch tactics, community growth." [E] An agency serving Series A-B SaaS is testing whether the depth is real for them or borrowed. [I]
**Purchase Decision:** They ask something narrow and vertical (a PLG-to-sales-assist motion, a category-creation sequencing question) and check whether the answer names SaaS-specific mechanics or slides into community and launch language. Internally: *"Is it deep in my world, or fluent in someone else's and improvising in mine?"*

### 4) Demonstrable advantage over the free tab: impacts whether this becomes a line item at all

**Importance:** 9/10. The objection is attested in a paying reviewer's own words in an adjacent category. [E]
**Explanation:** Every alternative claims practitioner training and compounding context in near-identical language, so comparison collapses to the incumbent already open on the desktop, where a reviewer found no "significant advantage over free alternatives like GPT or Gemini." [E]
**Purchase Decision:** They run the same prompt side by side in ChatGPT or Claude and in Hivemind, usually on a positioning question they already know the right answer to. Internally: *"Am I paying for a different answer, or a different login?"*

### 5) Scope coverage: impacts whether it handles the work clients are actually asking for on top of the retainer

**Importance:** 8/10. The scope is company-enumerated but unverified by any customer. [E]
**Explanation:** The unfunded strategic layer has a specific shape: positioning options, GTM sequencing, messaging frameworks, competitive differentiation, narrative development. [E] A tool that covers three of five leaves the founder's calendar as the bottleneck for the other two. [I]
**Purchase Decision:** They walk it through one real account end to end (positioning, then sequencing, then the messaging framework), rather than testing five prompts on five accounts. Internally: *"Can it carry a whole engagement's thinking, or just the easy first turn?"*
### 6) Time to first useful answer: impacts whether a pressure-test happens in the hour the angle is written

**Importance:** 8/10. No-onboarding is a stated product property, unverified. [E]
**Explanation:** The alternative senior opinion costs $150-500/hr or waits for a calendar block, which is why routine accounts never get one. [E] If the tool needs configuration before it is useful, it re-creates the delay it was bought to remove. [I]
**Purchase Decision:** They test it cold on a live brief within minutes of signup, before feeding it any context. Internally: *"Did I get something usable before I got invested in making it work?"*
### 7) Fit with where the work already lives: impacts whether the thinking reaches the deliverable or dies in a chat window

**Importance:** 8/10. The Notion-as-system-of-record behavior is verbatim from an agency describing its own stack. [E]
**Explanation:** These shops run on Notion, Google Docs, HubSpot, Asana. One states "Notion is the center of our content operation… Every article starts in Notion." [E] Hivemind has no integrations of any kind, so every output crosses that gap manually. [E]
**Purchase Decision:** They watch what they personally do with the third good answer, copy it somewhere, or leave it in the thread. Internally: *"Where does this actually go after I close the tab?"*

### 8) Output portability (Hidden gem): impacts the hidden per-engagement cost nobody prices at purchase

**Importance:** 8/10. The absence of export, templates, and document output is documented; the compounding cost is inferred. [E]
**Explanation:** There is no plan template, deck export, or structured deliverable. Output exists as conversation. [E] The second-order cost is that a founder who uses it on six accounts is now retyping and reformatting six times a month, exactly the "recurring tax in maintenance burden" pattern that quietly kills tools in this price band. It never appears in the trial, only in month three. [I]
**Purchase Decision:** They rarely test this deliberately. It surfaces when they try to turn a good session into a client-facing doc. Internally: *"Is this saving me the thinking, or just moving the typing?"*

### 9) (Off-mainstream but critical) Evidence that someone like them already bought it: impacts how much personal risk the founder absorbs alone

**Importance:** 8/10. The proof gap is verified as total across every source tier. [E]
**Explanation:** There is no named customer, case study, review-site presence, or quantified outcome anywhere. [E] At this ACV that is not a procurement blocker, but it means the founder is underwriting the bet with their own credibility, and this audience is already cynical that agencies adopt AI only "because other agencies are saying" they do. [E]
**Purchase Decision:** They search the vendor name, find name-collision products and no reviews, then ask a peer in a Slack group like Grow Your Agency or a Bureau room. Internally: *"If this turns out to be a wrapper, who else can I point at when my team asks why we bought it?"*

### 10) (Off-mainstream, deal-breaking) Client-material confidentiality (Hidden gem): impacts whether real discovery inputs ever get pasted in

**Importance:** 8/10. The vendor-side absence is verified; the trigger mechanism is attested in practitioner discussion. [E]
**Explanation:** No SOC 2, DPA, data-residency, or training-opt-out statement exists on any surface. [E] The constraint does not originate with the agency's own procurement, which is nonexistent at this price: it arrives through the client's AI policy, which agencies now report being asked to agree to as a contract clause. [E] The buyer is therefore enforcing someone else's compliance regime with no vendor documentation to satisfy it. [I]
**Purchase Decision:** They paste a sanitized version first, then look for a trust or security page and find none. Internally: *"If a client asks me in writing where their positioning brief went, what do I send them?"*

### 11) (Off-mainstream, high leverage) A counter-position the client's own assistant won't produce: impacts whether the founder frames the strategy call or reacts to it
**Importance:** 7/10, the repricing event is attested; the criterion itself is inference. [I]
**Explanation:** Clients now arrive with AI-formed positions ("I just got feedback from a client that is clearly from ChatGPT") and a general model gives both sides the same consensus answer. [E] The tool earns its keep only if it reaches an angle the client's assistant structurally cannot. [I]
**Purchase Decision:** They run the client's own likely prompt in ChatGPT, then run the harder version in Hivemind, and compare. Internally: *"When they push back with what their AI told them, do I have somewhere to go?"*

### 12) Pricing model fit, impacts whether it can be applied to every account or only the flagship

**Importance:** 7/10, tier structure is published, no price is, so this is inferred from the band. [I]
**Explanation:** Individual, team and agency plans exist with shared team context, but no price, trial terms, or seat caps are published anywhere. [E] The economics only work if the cost is low enough to apply speculatively, to pitch work and routine retainers, rather than being rationed like a fractional strategist. [I]
**Purchase Decision:** They divide the quoted price by active accounts and compare it to one hour of a fractional CMO. Internally: *"Can I put this on the $4K client, or does it only make sense on the one I'd have thought hard about anyway?"*
# Potential Partners

### 1) Notion
**How Used:** The system of record for the whole shop, covering client briefs, positioning docs, SOPs, and editorial calendars; one agency states plainly that every article starts in Notion. [E]
**When Used:** Every time a strategy decision needs to survive past the call that produced it, which is where the doc quietly stops being read. [E]
**Collab:** Publish a free "Positioning Pressure-Test" template in the Notion Marketplace: structured prompt blocks the founder pastes into Hivemind, plus a decision-log block to paste the pushback back into. Direction is manual and outbound-only, since Hivemind has no ingestion or export; the template *is* the integration and should be marketed as such. [I]

### 2) HubSpot (Solutions Partner ecosystem)
**How Used:** The CRM and automation hub for the agency's own pipeline and for most SaaS client accounts. [E]
**When Used:** During pitch prep and QBR season, when the founder needs an angle for an account already sitting in the CRM. [I]
**Collab:** Skip the App Marketplace. With no integration there is nothing to list. Instead co-run a session inside the Solutions Partner program: "Pressure-test your client's positioning live," Hivemind on screen, twenty partner agencies bringing real accounts. Artifact is the recording plus a partner-only trial code. [I]

### 3) Ahrefs
**How Used:** Keyword, competitive and backlink research at roughly $300/month combined with an on-page tool; the raw material every content retainer runs on. [E]
**When Used:** At the front of a content engagement, when competitive data has been gathered and someone still has to decide what it means. [I]
**Collab:** Sponsor Ahrefs Evolve (Oct 12–13, 2026, InterContinental San Diego, 600+ practitioners, single-track, practitioner-only) with a workshop titled "Your research is fine. Your angle isn't." Co-published artifact: a teardown series where Ahrefs supplies the data pull and Hivemind supplies the argument against the obvious read. [E]

### 4) Clay
**How Used:** Enrichment-heavy ABM, covering LinkedIn data, job-change signals, and funding triggers, for the agency's own outbound and for client outbound programs. [E]
**When Used:** When a target list is built and every account still gets the same generic first line. [I]
**Collab:** Build a Clay HTTP-API enrichment column that calls the Hivemind Knowledge API per account and returns a positioning angle. This is the one direction the product actually supports: other tools query Hivemind, Hivemind does not read them. Gate it behind the existing manual key issuance. [E]

### 5) Slack
**How Used:** Internal comms and client-facing Slack Connect channels, which can span up to 20 organizations. [E]
**When Used:** The 4pm "does this angle hold up?" message that currently goes to nobody senior. [I]
**Collab:** Ship a Slack app exposing a `/pressure-test` slash command against the Knowledge API, scoped to the published team plan's shared team context. Co-market through Slack's app directory listing plus a joint post on running an AI second opinion in a shared client channel without the client seeing the draft. [I]

### 6) OpenClaw / agent-skill registries
**How Used:** The AI-power-user slice of this segment already runs Claude Code, n8n and OpenClaw as a working stack. [E]
**When Used:** When an agency has built its own agent workflow and needs a marketing knowledge source it did not scrape itself. [I]
**Collab:** Harden the existing `myosin-hivemind` skill, currently a public repo with 1 star, 0 forks, 0 issues, into a self-serve key flow, then get it featured in the VoltAgent and LobeHub registries with a documented agency use case. Distribution artifact: a working `SKILL.md` an agency can fork in an afternoon. [E]

### 7) Grow Your Agency (Chris Bolton)
**How Used:** The founder's peer Slack; $35 lifetime, no subscription, 1,100–1,600 agency owners depending on which of the operator's own pages you read. [E]
**When Used:** When a founder is stuck on a pitch and wants a peer read before Monday. [I]
**Collab:** Sponsor a recurring weekly Community Counselor topic, "bring your hardest positioning question," where Hivemind's response is posted alongside three owners' human answers, unedited. Artifact: a monthly digest of the disagreements between them. [I]

### 8) (Hidden gem) Wynter
**How Used:** Message testing with verified B2B buyers, results back in under 48 hours, pay-as-you-go per test. [E]
**When Used:** The moment a founder has an angle they believe in and no way to defend it to a client who now has AI of their own. [I]
**Collab:** A bundled "angle to evidence" motion: Hivemind generates and pressure-tests three positioning options, Wynter tests them against real buyers, the founder presents human data with their name on it. This is the direct antidote to the slop risk, because the deliverable stops being model output. Artifact: a co-branded two-page results template the agency hands the client. [I]
### 9) (Hidden gem) G2
**How Used:** Where the founder and their SaaS clients check whether a tool is real. [E]
**When Used:** During the 2–4 week evaluation, at the exact point the "how is this different from ChatGPT" question gets asked. [E]
**Collab:** Get listed and reviewed, then exploit the mechanic almost nobody has priced in: since June 2026 G2 pipes verified reviews and Buyer Intent into ChatGPT, Claude, HubSpot Breeze, Gong and AirOps. Reviews are the only asset that answers the ChatGPT objection *inside ChatGPT*. Note the placement risk: G2's May 2026 "AI Marketing Agents" category describes autonomous campaign execution, which misdescribes this product. [E]

### 10) (Hidden gem) Sakas & Company
**How Used:** Karl Sakas's free monthly Agency Office Hours, second Tuesday, 12–1pm ET, three to four pre-selected questions at 10–15 minutes each; he has worked with 600+ agencies across 36 countries. [E]
**When Used:** When a founder has an operational or positioning problem they will not post publicly. [I]
**Collab:** Co-run one session per quarter where a submitted question gets Sakas's answer and Hivemind's, side by side, with Sakas free to say where the model is wrong. The format mirrors the product's own claimed mechanic, and the visible disagreement is the credibility asset. [I]

### 11) (Hidden gem) Setup® Agency Mastermind
**How Used:** Joe Koufman's Atlanta-run forums of six to eight non-competing agencies, scoped explicitly to 10–50 FTE marketing agencies, the tightest published match to this segment anywhere. [E]
**When Used:** Monthly moderated calls where founders bring their hardest commercial problems to peers. [E]
**Collab:** Offer forum-wide access as a member benefit in exchange for structured feedback: each forum runs one real client positioning question through Hivemind before the call and reports whether the pushback changed the answer. Artifact: the segment's first documented outcomes, which currently do not exist at any source tier. [E]

### 12) (Hidden gem) Sharebird
**How Used:** Free, attributed PMM Q&A where named practitioners answer positioning and messaging questions in public. [E]
**When Used:** When the founder or their client-side counterpart wants a specific answer from a named human rather than a framework. [I]
**Collab:** Sponsor a "practitioner vs. corpus" column and, more importantly, recruit contributors from it. The documented corpus is web3-scoped, and the SaaS depth the stated ICP requires has to come from somewhere nameable. Direction: contributors in, attribution out. [E]


# Communication Channels

### 1) Email (Google Workspace / Outlook), US

**What Used For:**
- Client-facing proposals, scopes, and pitch follow-ups; this inbox is where new business lives, so it is read fast and triaged faster [I]
- Vendor evaluation of anything under a few hundred dollars a month, where the founder is both evaluator and signer and no procurement thread exists [E, brief: founder-led sale, no procurement]

**How To Engage:**
- Open on the unfunded strategic layer in their own economics: the positioning work a client wants but won't fund, and the non-billable hours it eats [E: "clients opt to skip strategic planning in favor of immediate implementation"; "your agency does not reward non-billable activities"]
- Plain-text, from a named human at Myosin with an agency background, under 120 words, one paragraph plus one question; anything designed-looking reads as a sequence and gets archived [I]
- Ask for a reply, not a meeting: "what's the last positioning call you took that you wish someone had argued with you on?" [I]

**How To Apply:**
- Cold and re-engagement: lead with the adversarial demo transcript verbatim rather than a description of it, since the seller has no case study, customer logo, or review to cite anywhere [E, no proof assets found at any tier]
- Post-trial: a single check-in email asking which recommendation they pushed back on and whether it adapted, which is the one behavior the product claims as its own [E: "Tell it why a recommendation doesn't work and watch it adapt"]

**Tips:**
- Never write the subject line in the register they can spot; this buyer says they can identify AI copy "from a mile away" and hates it, and a slop-shaped cold email disqualifies a product sold on not being slop [E]
- Do not use the "$300/hr" or "without the agency markup" framing in writing to an agency founder; it argues against the fee they bill [E, seller's own copy]
- Send Tuesday to Thursday morning; Mondays are internal and Fridays are client escalations [I]
### 2) Slack, US

**What Used For:**
- Internal delivery traffic and account triage, plus shared Slack Connect channels with clients, which is where scope questions and "can you also do positioning" requests actually surface [E, Slack Connect used with clients, up to 20 orgs per channel]
- Peer chatter in paid agency-owner Slacks, where tools get recommended or dismissed in one line by someone the founder trusts more than any vendor [E, agency-owner Slack communities are paid and curated]

**How To Engage:**
- Enter peer Slacks only through a member who will vouch, and post an answer to someone else's positioning problem before ever naming the product [E: "it doesn't just turn into companies pitchslapping you" is the stated membership value]
- Format is a short reply inside an existing thread, three to five sentences, one concrete framework, no link on the first pass [I]
- The ask is a DM opt-in: "happy to run your actual client scenario through the thing that generated this, if useful" [I]

**How To Apply:**
- Pre-sale: treat peer-Slack credibility as the substitute for the missing review-site presence, since there is no G2 or Capterra listing for this product to point at [E]
- Post-sale: ask the founder to paste one pushback the tool gave them into their team channel. Adoption in a 10 to 50 person shop is social, and a strategy tool nobody sees dies quietly [I]

**Tips:**
- **(Hidden gem)** Assume anything you write to this founder can be pasted into a Slack Connect channel their client is sitting in. The agency's Slack is not a private perimeter, and a vendor line about replacing agency strategy becomes a client's negotiating quote the moment it is forwarded [I]
- Never post a launch announcement or a promo link in a peer Slack; these communities are explicitly built as a refuge from that, and one violation ends the channel [E]

### 3) Video calls (Zoom / Google Meet), US

**What Used For:**
- Client strategy sessions and new-business pitches, which is the exact setting where the founder gets asked why they are different from a cheaper alternative [E: "same inability to explain why his agency was different from cheaper alternatives"]
- Vendor demos, which at this price point are really trials with a person attached rather than evaluations [I]
**How To Engage:**
- Open by asking for a live client problem, not by sharing a slide: "give me a real account you're re-positioning this quarter and let's run it" [I]
- 30 minutes, camera on, screen shared for 20 of them, with the founder typing rather than the seller [I]
- Scheduling: send one link with two named slots in the same message. This buyer's own vendor uses "calendar ping-pong" as a pain phrase, so do not make them play it [E]

**How To Apply:**
- Discovery: use the call to qualify what no list can filter, namely whether they sell strategy alongside execution retainers and how many client brands they carry [E, no dataset segments agencies by client vertical]
- Close: the deal band is fast, roughly two to six weeks with one signer, so ask for the decision on the call rather than building a follow-up sequence [E, median cycle at this ACV is weeks, not quarters]

**Tips:**
- **(Hidden gem)** Let the founder try to make it agree with them, and run that test on purpose. Practitioners distrust models specifically because "sycophant mode has been a real problem of late," so a demo where the buyer wins every exchange is the demo you lose [E for the quote, I for the play]
- Do not demo the corpus as though it is SaaS-native; the documented knowledge base is web3-scoped, and a founder who asks for a B2B software analogue and gets a token-launch answer will not come back [E]
- Never present roadmap language such as autonomous execution or "AI CMO"; none of it is shipped [E]

### 4) LinkedIn, US

**What Used For:**
- Where this founder builds their own agency's authority, which means they read the feed as a competitor as much as a buyer [I]
- Where they vet a vendor before replying: they check whether real practitioners are behind it, which is the only claim this product actually makes [E, positioning rests entirely on practitioner provenance]

**How To Engage:**
- Open by commenting substantively on their post about a client outcome, days before any DM, so the profile they check is already familiar [I]
- DM format: four sentences, one specific reference to something they published, one screenshot of the product disagreeing with a marketer, no calendar link [I]
- Ask for a reaction to the screenshot, not a call: "would you have pushed back on that?" [I]

**How To Apply:**
- Named-practitioner posts from Myosin contributors carry the corpus claim better than the brand account, because the differentiator is who trained it [E]
- Mid-cycle: DMs are the right place to handle the free-alternative objection, since the founder is comparing against the ChatGPT tab already open on their desktop [E, proxy reviewer: no "significant advantage over free alternatives like GPT or Gemini"]

**Tips:**
- Publish the transcript, not the claim. Every competitor already says "not generic AI," so asserting it is parity noise, and only a visible argument differentiates [E, parity list]
- Do not use "agentic," "AI-powered," or "end-to-end" in a post; this vocabulary is now near-universal in agency-facing marketing and reads as filler to this reader [E]
### 5) Async video (Loom), US

**What Used For:**
- How agency teams hand off context internally and to clients without booking another call; Loom is already in the stack [E]
- How a founder shares a vendor find with a partner or ops lead without writing a summary [I]

**How To Engage:**
- Open with their own client's public positioning on screen inside the first eight seconds, no intro slide, no face-cam preamble [I]
- Three to four minutes, one take, ending mid-argument so the product's pushback is the last thing on screen [I]
- Ask them to reply by video or one line, whichever is faster, and say so explicitly [I]

**How To Apply:**
- Replaces the demo entirely for founders who will not book time during pitch season [I]
- Post-signup: a two-minute video showing how to phrase a prompt so the tool argues, which is the difference between activation and silent churn in a band where deals are lost to non-adoption rather than to competitors [E for the activation dynamic, I for the remedy]
**Tips:**
- Record against their portfolio page, not a generic example; a demo on someone else's brand proves nothing about their book [I]
- Do not script it word-perfect; polish reads as marketing, roughness reads as a practitioner [I]

### 6) In-person, US

**What Used For:**
- Small application-gated owner rooms and peer forums of six to eight non-competing agencies, where founders discuss pricing, margin, and what they are actually scared of [E, these formats exist and are capped small]
- Conference hallways and dinners, where tool recommendations pass founder-to-founder and outlast any ad [I]

**How To Engage:**
- Open with a margin question rather than a product one: what percentage of last quarter's strategy work was billed [E, 13% average after-tax margin makes this the live nerve]
- Format is one-to-one or a table of four; no booth energy, no deck, no swag [I]
- Ask to run one of their accounts live on a laptop right there, then hand them the keyboard [I]

**How To Apply:**
- Use the room to collect the proof asset the company does not have: one recorded, attributable win story with a named agency [E, zero named customers exist]
- Follow up within 24 hours by email referencing the specific account discussed, not the event [I]

**Tips:**
- Show the product arguing with you in front of them; describing it is parity, watching it is not [I]
- Do not pitch from a stage or sponsor a session in these rooms. The entry price is peer contribution, and vendors who buy their way in are visibly discounted [E]
# Sources

I'll verify each candidate source and URL before writing.
### 1) Agency Management Institute: Build a Better Agency (podcast, newsletter, and owner peer networks; Kansas-based, US small-to-mid independent agencies), the only mainstream media property whose entire audience definition *is* this persona: the owner of an independent 10–50 person shop.
**Content:** Weekly 30–45 minute owner-level episodes plus AMI's research and owner-network content, aimed squarely at agency economics: niching, pricing, retainer structure, non-billable drag. This founder listens on the drive between pitches and reuses the frames verbatim in partner conversations about what to stop doing for free. [E]
**How To Apply:** Pitch a guest episode built on the one thing Hivemind can prove on screen: an adversarial strategy session where the model refuses the brief and asks for the retention loop first. Then sponsor the BABA Summit (May 24–26, 2027, The Westin Denver Downtown, registration live) rather than the feed, because the room is owners only. [I]
**Link:** https://agencymanagementinstitute.com/babasummit/

### 2) HubSpot Partner ecosystem (Solutions Partner Program and Solutions Directory; Cambridge, MA, global), the platform this segment's revenue actually runs through, and the ecosystem that decides which agencies get seen.
**Content:** Tier requirements, certification tracks, co-marketing mechanics, and directory placement rules. Founders read it as a lead-source and a credential; the dossier shows HubSpot as the CRM hub in agencies publishing their own B2B SaaS stack. [E]
**How To Apply:** Hivemind has no HubSpot integration and should not imply one. Contribute a partner-blog-style asset instead: a positioning pressure-test worksheet for HubSpot partners pitching SaaS clients, output produced live in one Hivemind thread from angle through landing page headline. [E]
**Link:** https://www.hubspot.com/partners

### 3) CXL (practitioner marketing training and certification; Austin, TX), where this founder sends themselves and their two most senior people when the gap is judgment, not tooling.
**Content:** Minidegrees and single courses taught by named practitioners, including a marketing strategy certification track and B2B programs. Founders buy seats to make a mid-level strategist safe to put in front of a client, which is the same substitution Hivemind is proposing at a lower price point. [E]
**How To Apply:** Sponsor or co-author a CXL lesson module on stress-testing a positioning hypothesis, with the Hivemind transcript as the artifact students critique. The pitch to CXL is content, not a tool trial: their audience already accepts AI for strategy and operations while refusing it for creative. [E]
**Link:** https://cxl.com/institute/online-course/marketing-strategy/

### 4) Wynter (B2B message testing with verified buyer panels; Estonia/US, founded by Peep Laja), the segment's reference point for what counts as *evidence* that a message lands, which is exactly the ceiling Hivemind's output will be held to.
**Content:** Message, pricing-page, and preference tests returned in under 48 hours from panels of real target buyers, sold pay-as-you-go. Agency founders use it to convert a strategy recommendation into something defensible in front of a client instead of an opinion. [E]
**How To Apply:** Co-created asset, not competition: run one Hivemind-generated positioning angle against the client's incumbent angle through a Wynter test and publish the paired result. This is the cleanest available answer to "it's a benchmark, not a source of truth," and Hivemind has no measurement capability of its own to make the claim unaided. [E]
**Link:** https://wynter.com/products/message-testing

### 5) Grow Your Agency Slack (agency-owner peer community, Chris Bolton; US-anchored, remote), a low-friction, high-trust room of owners at exactly this size, with no subscription to justify to a partner.
**Content:** Weekly curated discussion prompts, an owner directory, and unguarded threads on pricing, scope, and what founders are actually doing with AI. Karl Sakas calls it the internet's best no-subscription community for agencies. [E]
**How To Apply:** Buy in as a member-operator, not a sponsor, and run a recurring "bring your worst positioning problem" thread where Hivemind's pushback is posted publicly with the founder's rebuttal. The product's only differentiated behavior is being argued with, which means the demo has to happen in front of peers. [I]
**Link:** https://growyouragency.group/agency-community/
### 6) Fletch PMM resource library (B2B positioning and messaging templates, Anthony Pierri and Rob Kaminski; US), the segment's default free scaffolding for B2B software positioning, and the human benchmark for output quality.
**Content:** Public positioning templates, a documented positioning and messaging process, and hundreds of homepage teardowns, from a firm claiming 500+ B2B software companies helped via two-week sprints. Founders lift the framework and run it themselves on a client who cannot fund a sprint. [E]
**How To Apply:** Publish a companion piece that takes a completed Fletch-style positioning canvas and shows Hivemind interrogating it: which claim is undefended, which competitor also says it. Positioning as the layer *after* the template, never as its replacement. Note the parity trap: "not generic frameworks" is already competitor-standard language and cannot carry the claim. [E]
**Link:** https://www.fletchpmm.com/resource/startup-positioning-template

### 7) G2 (software review marketplace and buyer-intent data; Chicago, IL), the surface where the ChatGPT comparison gets litigated by strangers, and now inside the assistants this founder already uses.
**Content:** Verified reviews, category grids, and buyer-intent signals, with new AI categories added through 2026 including AI marketing agents. G2 reviews now feed ChatGPT, Claude, and adjacent tools, so a founder asking an assistant for alternatives is served G2 data. [E]
**How To Apply:** Hivemind has zero review presence at any tier, which is the actual bottleneck, not the messaging. Seed the first fifteen reviews from agency-tier accounts and fight the category assignment: landing in an execution-automation category misdescribes an advisory product and invites the wrong comparison. [E]
**Link:** https://company.g2.com/news/new-categories-introduced-in-may-2026

### 8) Clutch (B2B services ratings and reviews marketplace; Washington, DC), the services marketplace that both supplies this founder's inbound and makes the segment enumerable for targeting.
**Content:** Verified firm profiles, client interviews, and ranked category lists including B2B marketing agencies, with a published research methodology. Founders maintain their profile as a lead channel; agency owners on record report steady client flow from it. [E]
**How To Apply:** Use Clutch as the list-build layer first, filtering US firms in the 10-50 band and hand-verifying majority-B2B-software rosters from portfolio pages, since no dataset segments agencies by client vertical. Then publish a teardown series on the differentiation language in Clutch profiles, which is the exact wound: unable to say why you beat a cheaper alternative. [E]
**Link:** https://clutch.co/agencies/business-services-industry

### 9) Digiday (marketing and agency trade press; New York, NY), the outlet that sets what agency leaders believe is happening to AI in their industry this quarter, including the governance questions clients now ask.
**Content:** Reported pieces on agency AI pitches facing harder scrutiny, data ownership and portability moving from procurement into governance, and skepticism toward AI tools priced above their proof. This founder forwards these to partners as ammunition for client conversations. [E]
**How To Apply:** Do not pitch a product story; there is no customer, case study, or outcome number to hang it on. Offer the founder as a source on the inversion Digiday has not covered: clients arriving with AI-generated feedback, and what that does to the price of judgment. [I]
**Link:** https://digiday.com/marketing/agency-ai-pitches-are-starting-to-face-harder-questions/

### 10) (NON-MAINSTREAM HIDDEN GEM) Sakas & Company Agency Office Hours (free live monthly Q&A, Karl Sakas, Raleigh, NC), a public feed of real founder questions from an advisor who has worked with 600+ agencies across 36 countries, recurring the second Tuesday monthly at noon ET; next session recurring, no annual edition to verify.
**Content:** Three or four pre-selected questions answered at 10-15 minutes each, with recaps published as posts on pricing, retainers, revenue plateaus, and USP. The questions are unpolished founder problems, which makes this the highest-signal source of *live* objection language for this persona anywhere in the report. [E]
**How To Apply:** Attend, log the recurring questions verbatim, and build the gated offer directly from them: a "pressure-test your USP" session where Hivemind argues against the founder's own differentiation claim. Sponsorship is unlisted and probably unavailable, so treat this as intelligence and message-source, and route the audience to your own asset. [I]
**Link:** https://sakasandcompany.com/faq-agency-office-hours-with-karl-sakas/

# Reach Channels

I'll verify the events and communities before writing.
### 1) LinkedIn organic, founder-to-founder posting (personal profile of the seller's founder, not the company page): *Trigger, Self-education*
**Relevance:** Agency founders live on LinkedIn because it is where new business originates, and they read positioning content from Fletch PMM, April Dunford, and Emily Kramer daily. They post about client work and pitch wins, so their trigger moments are publicly legible before they ever search for a tool. `Best practices:` Fletch PMM posts annotated homepage teardowns as image carousels with the "before" screenshot as slide one; Emily Kramer (MKT1) publishes frameworks as downloadable artifacts and routes them through her own profile rather than a brand page. [E]
**How To Use:**
- Weekly carousel: screenshot a real prompt asking for a full launch plan, screenshot Hivemind refusing and asking "what's your retention loop?", caption "Your AI should be doing this to your strategy before your client does." CTA: comment "PUSHBACK" for the full transcript. [E]
- Series titled *Non-Billable Hours*, one post per week naming a specific strategy task founders absorb unpaid (pitch positioning, QBR narrative, retainer re-scoping) and what a second opinion that pushes back changes about it. CTA: DM for a free pressure-test on one live account. [I]
- Comment-first campaign: reply substantively under Fletch and Dunford positioning posts with a contrarian angle, never a link, to build recognition before the first cold DM lands. [I]

### 2) Search: artifact-level SEO and paid, not category-level, *Self-education, Shortlist*
**Relevance:** There is no settled category name, and G2's nearest 2026 addition, "AI marketing agents," describes campaign automation and misdescribes this product. Buyers therefore type job tasks: "AI positioning tool," "ICP generator," "AI marketing brief tool," "how to pressure-test positioning." That demand is capturable today without funding a category. `Best practices:` M1-Project ranks by publishing one page per artifact (ICP generator, persona generator) rather than one page for the platform; Fletch PMM ranks on positioning queries by giving away the Figma and Notion templates the query is looking for. [E]
**How To Use:**
- Build four artifact pages (positioning options, messaging framework, GTM sequencing, competitive differentiation), each ending in a live chat that opens with a challenge question rather than a form. [E]
- Publish `hivemind vs chatgpt for marketing strategy` as an honest comparison that concedes ChatGPT wins on cost and switching, and claims only the adversarial mechanic. CTA: run the same prompt in both, side by side. [I]
- Paid search on `fractional CMO cost` and `positioning consultant pricing`, landing on a page that frames the product as the pre-work you do before you spend $4k-20k/mo on a human. [E]
### 3) Newsletter sponsorship, agency-owner titles first, *Self-education, Shortlist*
**Relevance:** Agency-founder newsletters are small (BHIVE's ~6,000 subscribers is typical) but titled precisely; the large B2B marketing lists sit on the in-house side. Sponsorship exists at Exit Five, MKT1, and Why We Buy with a named contact route, but none publishes a rate card, so pricing is negotiated. `Best practices:` Why We Buy runs sponsor copy in the host's voice after five short follow-up questions rather than accepting supplied creative; Exit Five sells across newsletter, podcast, and LinkedIn as one bundle. [E]
**How To Use:**
- Sponsor Sakas & Company's semi-weekly agency newsletter with a single-line ad reading "The senior second opinion you can't bill for," linking to a page of ten real pushback transcripts. [E]
- Buy MKT1 placement (sponsorships@mkt1.co) positioned to the AI-native-marketing-function thesis Kramer is already writing against in 2026. [E]
- Run a Why We Buy native placement built from their five-question process, with the hook "the client has ChatGPT now too" and a CTA to a free pitch-prep session. [I]
### 4) Webinar / live teardown: *Shortlist, Evaluation*
**Relevance:** This buyer will not sit through a product demo but will watch a peer's positioning get taken apart, because that is the exact skill they are buying. The format also solves the proof gap: with zero case studies, named customers, or reviews anywhere, live unrehearsed output is the only evidence available. `Best practices:` Wynter runs message tests and shows the raw buyer verbatims on screen; Fletch PMM built its reputation on public teardowns where the flawed original is shown first. [E]
**How To Use:**
- Monthly *Pressure-Test Live*: three agency founders submit a real client's positioning, Hivemind challenges it on screen, a human strategist adjudicates. CTA: submit yours for next month. [I]
- Co-host with a mastermind operator (Grow Your Agency, Setup®) so the room arrives pre-vetted and the host's endorsement carries. [E]
- *Pitch Rehearsal* clinic in the two weeks before a known RFP season: bring a prospect you're chasing, leave with three angles and the objection you hadn't planned for. [I]

### 5) Video, screen-recorded transcript walkthroughs, YouTube and embedded: *Self-education, Evaluation*
**Relevance:** The differentiating claim is a behavior, and behavior does not survive a bullet list. Founders evaluating on "will it argue or will it agree" need to watch it argue. Sycophancy is a named, live complaint about current tools in this audience, so the video must include the moment the model refuses the user's premise. `Best practices:` Tas Bober built an audience on 400+ website teardowns recorded as short screen captures; Ahrefs wins by making the tool itself the on-screen subject rather than a talking head. [E]
**How To Use:**
- Six-minute *Watch It Disagree With Me* series: founder states a positioning they believe, the tool objects, founder pushes back, tool adapts. No cuts. [E]
- 45-second vertical cuts of the single hardest question asked in each session, captioned, posted to LinkedIn and Shorts. CTA: full session in the first comment. [I]
- A "compare it yourself" video where the same prompt runs in ChatGPT and Hivemind in split screen. The objection is going to be litigated anyway, so litigate it on camera. [E]

### 6) Communities, Slack rooms where agency owners already talk: *Trigger, Shortlist*
**Relevance:** Grow Your Agency's Slack is a one-time $35 lifetime join with a stated 1,100-1,600 owners and weekly curated discussion prompts, which makes it the cheapest room in the segment to be genuinely present in. Online Geniuses carries agency owners at scale but with far lower title purity. `Best practices:` GrowthMentor earns placement by publishing the definitive roundup of the communities it competes with; Exit Five members publicly police pitching, which is why non-promotional presence converts there. [E]
**How To Use:**
- Join Grow Your Agency and answer the weekly Community Counselor prompt with real strategy reasoning for eight weeks before mentioning the product once. [E]
- Offer the community a free members-only pressure-test week, coordinated with Chris Bolton, positioned as a perk from the operator rather than a vendor promotion. [I]
- Seed a thread in Exit Five's AI channels on where members stop trusting AI output, contributing the permission structure this audience already uses: back-of-house strategy yes, client-facing creative no. [E]
### 7) Private networks, peer masterminds and small-room retreats: *Trigger, Decision & rollout*
**Relevance:** Setup® Agency Mastermind runs forums of 6-8 non-competing agencies scoped explicitly to 10-50 FTE marketing shops, which is the segment definition verbatim. The Bureau's Owner Camps cap at roughly 20 seats. These are the rooms where founders admit what they can't bill for. `Best practices:` Setup® gates by non-competition so members share operating detail freely; The Bureau qualifies after deposit, which is why its alumni recommend tools to each other by name. [E] **(Hidden gem)** Non-competition is the distribution mechanic. In a room of agencies that will never pitch against each other, there is no incentive to conceal a tool advantage, so a single adopter's endorsement propagates to seven peers instead of dying as a secret. Sponsor the forum, not the individual. [I]
**How To Use:**
- Offer Joe Koufman a free forum-wide trial for one Setup® cohort in exchange for a facilitated debrief session at month two. [E]
- Sponsor a Bureau Leadership Love or Owner Camp session slot and run it as a working session on pricing strategy work, not a talk. [E]
- Build a referral term specifically for mastermind members: one free seat for the referrer's own account for each peer-forum member who converts. [I]
### 8) Podcasts: agency-owner shows, guest and sponsor, *Trigger, Self-education*
**Relevance:** 2Bobs (Baker and Enns, ~231 episodes) and Build a Better Agency (Drew McLellan) are the two shows this segment cites by name, and both already run episodes on whether AI changes agency positioning. Smart Agency Masterclass has ~900 episodes aimed at seven- and eight-figure owners. `Best practices:` 2Bobs sustains authority by arguing on air rather than interviewing agreeably; Product Marketing Life openly sells episode sponsorship, making it the low-friction test buy. [E] **(Hidden gem)** A listener review of 2Bobs states he picks an episode before every new-prospect meeting. That means the ad is heard in the pitch-prep window, not the browsing window, so creative should sell pitch confidence, never seat efficiency. [E]
**How To Use:**
- Pitch Drew McLellan an episode titled "Why your clients stopped paying for strategy" built on the billable-versus-non-billable frame his audience already uses. [E]
- Buy a mid-roll on Smart Agency Masterclass with copy addressed to the drive to a pitch: "Before you walk in, have something argue with your angle." [I]
- Sponsor Product Marketing Life as the cheap A/B against agency shows to test whether the secondary in-house buyer responds to the same mechanic. [E]

### 9) Events: booth-free, session-and-hallway motion, *Shortlist, Evaluation*
**Relevance:** Four dated, verified windows sit ahead of today: AgencyCon, October 8-9 2026, Children's Museum of Denver; MAICON, October 13-15 2026, Cleveland; MarketingProfs B2B Forum, November 2-4 2026, Boston; AMI Build a Better Agency Summit, May 24-26 2027, Westin Denver Downtown. Exit Five Drive runs September 8-10 2026 at Spruce Peak, Stowe, for roughly 400 attendees and reaches the secondary buyer. `Best practices:` Exit Five programs sessions on where marketers stop trusting AI output rather than on tools; AgencyCon keeps the room to agency leaders only, so hallway conversation stays on-topic. [E]
**How To Use:**
- Submit an AgencyCon or BABA session titled "Pricing the strategy you're currently giving away," with the product appearing only as the live demo in the final ten minutes. [I]
- Run an invite-only breakfast for fifteen founders adjacent to MAICON, structured as three live pressure-tests of attendees' own client positioning. [I]
- At MarketingProfs B2B Forum, skip the booth and record hallway pushback sessions on video for the Q1 content calendar. [I]

### 10) Platform-native: the answer-engine and review-graph surface (G2 plus AI-assistant citations), *Self-education, Shortlist*
**Relevance:** As of June 2026 G2 pipes verified reviews and buyer intent into ChatGPT, Claude, HubSpot Breeze, Gong, Profound, and AirOps. This buyer researches alternatives inside the exact assistant they are comparing you to, which means the ChatGPT objection is now litigated using G2 data with no Hivemind listing present. Product Hunt is the other native surface; the launch already happened there. `Best practices:` AirOps and Profound wired themselves into the review graph early and are cited by name inside those assistants; M1-Project publishes reviewer verbatims directly on its agency page. [E]
**How To Use:**
- Create the G2 listing this week and run a fifteen-customer review drive, since a listing with zero reviews is the current state and is worse than a thin one. [E]
- Publish a structured, crawlable comparison page, covering mechanism, corpus scope, and what it does not do, written to be quoted verbatim by an assistant answering "AI marketing strategy tools for agencies." [I]
- Re-engage the Product Hunt comment thread and convert any real user language there into the first attributable proof asset the product has. [E]
# Sources: Jobs To Be Done

I'll verify sources for each of the six jobs before writing.
All sources verified. Writing the appendix.

## Market Practices Behind Selling and Delivering the Strategic Layer at Boutique B2B SaaS Agencies (2022-2026)

### Fact 1: Agencies fund a point of view about the prospect's market out of their own pocket, because the pitch is where positioning work is judged and nowhere is it paid for

**Summary:** Speculative strategic work in new-business pitches is a costed, unreimbursed norm (enough so that the industry's own trade bodies publish guidance on stipends and on who owns the thinking presented). Meanwhile the positioning work itself is priced as a discrete product elsewhere in the market, at consultancy rates, which is why it is affordable for a signed client and not for a prospect. The two facts together describe the practice: agencies show a point of view they could not justify buying. [E]

**Sources:**
- **10 Positive Pitch Principles** — 4As (American Association of Advertising Agencies) — https://www.aaaa.org/blog/ana-and-4as-10-principles-of-pitching/
- **Pricing** — Fletch PMM — https://www.fletchpmm.com/pricing

**Evidence:** The 4As/ANA principles state agencies "incur hard costs for every pitch" and recommend stipends plus agency retention of IP in presented work, which establishes speculative strategy as priced-but-unpaid. Fletch PMM's own site sells positioning to B2B software companies as a two-week sprint, showing the same output has a market price the moment a buyer commits, and none before. [E]

**Key phrases:** agency pitch stipend, 4As ANA positive pitch principles, speculative creative IP ownership, positioning sprint pricing, new business point of view

---

### Fact 2: Strategy is absorbed into execution retainers as unfunded senior work, and the only documented fix requires unbundling it into a separately sold engagement

**Summary:** Average agency economics leave almost no room for senior hours that carry no invoice, so positioning and GTM work delivered inside a retainer lands directly on owner profit or owner evenings. The advisory market's answer is Paid Discovery: price the strategic phase as its own engagement before any implementation scope is signed. Reconciling this with the unfunded-layer thesis: Paid Discovery is evidence that strategy *becomes* priceable only when deliberately sold as a standalone product to a client who agreed to buy it. The quoted range sits at the high end of the advisory market and does not describe the routine account or the pitch, where the same work is expected inside an existing scope and remains unfunded. [E]

**Sources:**
- **How Profitable are Digital Agencies?** — Promethean Research — https://prometheanresearch.com/how-profitable-are-digital-agencies/
- **Get started with Paid Discovery at your agency: Implementation tips** — Sakas & Company — https://sakasandcompany.com/start-using-paid-discovery/

**Evidence:** Promethean reports the average digital agency earned a 13% after-tax net margin in 2025, below its long-run average since 2015, quantifying how little slack exists for non-billable senior time. Sakas documents the structural countermeasure and the advisory-market prices attached to it, which shows strategy is sellable when unbundled and confirms it is unfunded whenever it is not. [E]

**Key phrases:** paid discovery agency, agency net margin benchmark, non-billable senior hours, unbundling strategy from retainer, Promethean State of Digital Services

---

### Fact 3: Owner-operators buy structured external challenge in small paid rooms, while the assistants already on their desktop are documented to default to agreement

**Summary:** Peer-advisory forums for agency owners exist and charge for one specific good: a group of non-competing operators willing to contradict the person who signs the invoices, on a monthly cadence. The model vendors have separately published and rolled back over-agreeableness in consumer assistants. Both describe the same market condition this segment operates in: contradiction is a purchased service, and the free tool is engineered toward approval. [E]

**Sources:**
- **Agency Mastermind**, Setup®, https://setup.us/agencymastermind
- **Sycophancy in GPT-4o: What happened and what we're doing about it**, OpenAI, https://openai.com/index/sycophancy-in-gpt-4o/

**Evidence:** Setup® structures its program as forums of six to eight non-competing agencies scoped to 10–50 employee marketing agencies with monthly moderated calls, establishing paid peer challenge as an existing category serving exactly this firmographic. OpenAI's post confirms the counterpart by describing the withdrawn update as "overly flattering or agreeable," the failure mode practitioners in this segment already name and distrust. [E]

**Key phrases:** agency owner mastermind non-competing, executive isolation decision quality, LLM sycophancy rollback, monthly peer forum agency, devil's advocate strategic review

---
### Fact 4: Differentiation is treated as a live, repeatedly defended argument rather than a completed deliverable, and consultancies sell specifically against the dormant-deck failure

**Summary:** The positioning-advisory market for agencies is built on a named failure pattern: the firm has done the exercise, owns the deck, and still cannot answer why it costs more than a cheaper option. Agency-facing consultants and pitch advisors both sell the same remedy: replacing a static artifact with a stated position the principal can defend live in a room. The practice being observed is not producing positioning; it is rehearsing and holding it under commercial pressure. [E]

**Sources:**
- **Agency Positioning: How to Choose Who You Serve and Why It Changes Everything**, Haus Advisors, https://www.hausadvisors.com/blog/agency-positioning
- **Stop Pitching Your Agency in 2026 and Start Proving Your Strategic Value**, Business of Story, https://businessofstory.com/ad-agency-new-business-pitching/

**Evidence:** Haus Advisors' case narrative describes an agency founder who had "already done the positioning work" and "had a deck" yet showed the "same inability to explain why his agency was different from cheaper alternatives," naming the failure precisely (note: this is a rival consultancy's own marketing, not neutral research); Business of Story reaches the same conclusion from the pitch side, quoting an agency principal that building custom brand strategy requires "tens of thousands of dollars in internal resources and tons of time," which explains why the argument is under-rehearsed. [E]

**Key phrases:** agency positioning dormant deck, differentiate from cheaper alternatives, proving strategic value pitch, cosmetic positioning agency, defending agency rate

---

### Fact 5: Sameness, not error, is the recognized quality failure in AI-assisted marketing work, and the language of agency AI marketing has already converged

**Summary:** The convergence of AI-assisted output is now a mainstream, named phenomenon: the word for undifferentiated AI content entered dictionary-level general usage in 2025. In agency marketing specifically, reviewers of platform positioning report the descriptions have become interchangeable enough that authorship is unidentifiable once branding is removed. Agencies therefore manage a documented sameness risk in how their work reads, and buyers have begun asking harder questions about it. [E]

**Sources:**
- **Word of the Year 2025: Slop**, Merriam-Webster, https://www.merriam-webster.com/wordplay/word-of-the-year
- **Agency AI pitches are starting to face harder questions**, Digiday, https://digiday.com/marketing/agency-ai-pitches-are-starting-to-face-harder-questions/

**Evidence:** Merriam-Webster's selection of "slop" establishes that undifferentiated AI output has a settled name in general usage, which is also the word practitioners in this segment use unprompted; Digiday reports a platform review finding the descriptions so uniform that "stripped of branding, it would be nearly impossible to attribute" one to its source (scope caveat: that review concerns holding-company platforms, not boutique tools, so treat it as evidence about category language rather than about boutique buyer speech). [E]
### Fact 6: Fee conversations have shifted onto AI-assisted production cost, and agencies are being advised not to price against it

**Summary:** Clients now reason that AI lowered the cost of the work and open scope conversations from there, which turns a value discussion into a cost-plus one. Agency advisors are publishing explicit counter-positions: do not discount AI-assisted work, because doing so trains clients to price the output rather than the judgment. The structural warning underneath is that a cost argument is unwinnable for a boutique. [E]

**Sources:**
- **The Biggest Threat To Agencies Isn't AI; It's Commoditization,** Forbes Agency Council, https://www.forbes.com/councils/forbesagencycouncil/2025/03/11/the-biggest-threat-to-agencies-isnt-ai-its-commoditization/
- **How Agencies Are Pricing AI-Assisted Work in 2026,** Kampaign Lab, https://www.kampaignlab.com/agency-freelance/how-agencies-are-pricing-ai-assisted-work-in-2026

**Evidence:** The Forbes Agency Council piece, authored by a named agency principal, states plainly that "you can't win a price war with someone whose costs are near zero," establishing why cost-based defense fails for this segment. Kampaign Lab documents the working counter-frame (that nobody expected a discount because a photographer used autofocus, and that discounting AI-assisted work trains clients to devalue it), which is practitioner advice rather than measured buyer behavior and should be cited as such. [E]

**Key phrases:** clients expect AI discount agency fees, agency commoditization threat, pricing AI-assisted work, value versus cost-plus scope, agency rate erosion
# Sources: Goals Or Objectives

## Market Expectations for the Unfunded Strategic Layer at Boutique B2B SaaS Agencies (2022–2026)

### Fact 1

**Summary:** Strategic planning at small agencies is structurally unfunded on both sides of the transaction. Clients deprioritize it in favor of implementation, and the agency's own comp and utilization model gives no credit for the hours it consumes. It lands on the founder's unbilled time inside a business already running thin margins. [E]

**Sources**
- Promethean Research, *How Profitable are Digital Agencies?*, https://prometheanresearch.com/how-profitable-are-digital-agencies/
- Jeff Meade, *Why It's So Hard to Sell Strategy* (Marketing Agency Playbook), https://meade.substack.com/p/why-its-so-hard-to-sell-strategy

**Evidence**
- Promethean reports an average after-tax net margin of **13% in 2025** against a long-run figure of roughly **15%** since 2015, on an average sample revenue of **$4.43M**. Implied after-tax profit on that average agency is around **$575k**: the pool every unbilled senior hour is drawn from. [E]
- Meade, writing to agency owners, states that clients "opt to skip strategic planning in favor of immediate implementation," a demand-side problem, not a delivery failure. His companion piece on unbillable work states plainly that "your agency does not reward non-billable activities." [E]
- The same corpus supplies the internal cost figure from an agency principal: developing custom brand strategy "requires tens of thousands of dollars in internal resources and tons of time." That is why speculative strategy is rationed to flagship accounts rather than applied across the book. [E]
- **Verification limit:** Promethean's per-FTE and pricing benchmarks sit behind the paid report. Revenue-per-FTE of ~$143k used elsewhere in this report is derived from $4.43M ÷ 31 FTE, not published. Do not present it as sourced. [I]

**Key phrases:** non-billable, billable vs. non-billable, scope, retainer, net margin, "skip strategic planning," internal resources

---

### Fact 2

**Summary:** The 10–50 FTE founder-led agency concentrates buying authority and strategic authorship in one person, which makes the sale fast and the output capacity narrow. Delegation of strategy demonstrably happens at this size; the constraint is the founder's review, not permission. [E]

**Sources**
- Promethean Research, *2026 State of Digital Services*, https://prometheanresearch.com/2026-state-of-digital-services-digital-agency-industry-research/
- ProperExpression, *B2B Marketing Account Manager (Agency)*, remote US, https://builtin.com/job/b2b-marketing-account-manager-agency/7003350

**Evidence**
- Promethean's 2026 sample: **n=119** agency owners/managers, primarily North America, **74% United States**, average **31 employees**, average revenue **$4.43M**, and **68% founders, owners or partners**. Read the 68% as sample composition, not a population estimate. It still establishes that the person benchmarking the business is the person who would sign a tool invoice. [E]
- Promethean's wider count puts **>50,000 digital agencies in US and Canada** with **88% under 10 FTE**, placing the 10–50 band in roughly the top 12% of the population. This is a narrow enumerable list, not a long tail. [E]
- The ProperExpression posting is live agency demand for strategic delegation, recruiting against the pain directly: candidates "tired of managing 17 accounts and getting no strategic seat at the table," with the counter-offer that "our AMs lead strategy." Agencies this size are hiring to move strategy off the founder. [E]
- Consequence for positioning: framing the product as "you can't delegate strategy" contradicts observable hiring behavior. The defensible frame is that delegated strategy still routes back through the founder for judgment. [I]

**Key phrases:** founder-led, owner/partner, account manager, strategic seat at the table, headcount band, no procurement

---

### Fact 3

**Summary:** This buyer can identify AI-generated output on sight and treats generic AI text as a category of contempt, not a quality complaint. The same practitioners grant themselves explicit permission to use AI on strategy and operations while refusing it on client-facing craft. That split is the operative permission structure. [E]

**Sources**
- Glassdoor Community, Advertising bowl, *"Are you guys putting your decks into ChatGPT? What are you learning"* (Dec 8, 2025), https://www.glassdoor.com.au/Community/advertising/are-you-guys-putting-your-decks-into-chatgpt-what-are-you-learning
- Merriam-Webster, *Word of the Year 2025: Slop*, https://www.merriam-webster.com/wordplay/word-of-the-year

**Evidence**
- In the Glassdoor thread, a practitioner dismisses the tool as "a computer that just spits out aggregated slop," while another draws the boundary the product must respect: "Yes but mostly for strategy or operations, never for creative." Both registers appear in the same conversation. [E]
- The same thread surfaces the confidentiality constraint without being asked: decks contain proprietary client information, and clients' own AI policies now carry clauses agencies are asked to accept. Any pitch that increases the volume of client material passing through a third-party model must answer this. [E]
- "Slop" became Merriam-Webster's 2025 word of the year, confirming the term is general-register rather than niche practitioner slang. Use the buyer's word; do not translate it into "low-quality AI output." [E]
- The sharpest first-person version of the recognition claim, "I can spot AI-generated copy from a mile away. And I hate it… It's all starting to sound the exact same," comes from an agency-owner newsletter. **Verification limit:** only the publication root resolved on re-check (https://multipletabsopen.substack.com/); the specific post URL did not. Treat the quote as dossier-sourced, not independently re-verified. [E]
- Note the register: the fear is not inaccuracy. It is fluffiness, agreeableness, and interchangeability. Accuracy claims will not land. [I]

**Key phrases:** slop, generic, fluffy, on-brand, brand voice, back-of-house, "never for creative," AI policy clause
### Fact 4

**Summary:** Differentiation failure at this segment size is a documented, recurring commercial wound, and completing a positioning exercise does not resolve it. Selling strategy requires making the cost of its absence tangible, a sales problem the agency has to solve for itself before it can solve it for clients. [E]

**Sources**
- Haus Advisors, *Agency Positioning: How to Choose Who You Serve and Why It Changes Everything*, https://www.hausadvisors.com/blog/agency-positioning
- Sun Yi, *Selling The Invisible: How To Position Strategy So Clients Actually Buy*, Forbes Agency Council (May 7, 2026), https://www.forbes.com/councils/forbesagencycouncil/2026/05/07/selling-the-invisible-how-to-position-strategy-so-clients-actually-buy/

**Evidence**
- The Haus case narrative describes a founder who had "already done the positioning work," had hired a brand strategist, received exercises, ICPs, value props, a deck and a tagline, and six months later showed "the same inability to explain why his agency was different from cheaper alternatives." The artifact existed. The capability did not. [E]
- **Source caution:** Haus Advisors is itself a positioning consultancy selling against exactly this failure. The pattern is credible, and the diagnosis is commercially motivated. Never present this narrative as neutral third-party research. [E]
- The Forbes Agency Council piece names the mechanic for selling an intangible: "You make the cost of not having it tangible," and frames the work as "better positioning, better questions and better restraint." That is a pushback-and-questions frame, not a deliverable-volume frame, and it aligns with the adversarial mechanic this product claims. [E]
- Priced comparison the buyer will make: Fletch PMM sells the same output as a named human authority. "We've helped 500+ B2B software companies clarify their messaging and homepage in 2-week positioning sprints" (https://www.fletchpmm.com/). The tool does not beat that on authority. It beats it on being available speculatively, pre-sale, across a whole client book. [E]

**Key phrases:** positioning, ICP, value prop, differentiation, cheaper alternative, pitch, new business, "cost of not having it"

---

### Fact 5

**Summary:** The asymmetry has already flipped. Clients now run agency work through their own models and return AI-generated feedback, which reprices judgment from the buy side. Combined with near-zero-cost production, this makes output volume an indefensible thing to sell and pushback the only remaining product. [E]

**Sources**
- Glassdoor Community, Advertising bowl, *"I just got feedback from a client that is clearly from ChatGPT. They uploaded the work and had it give notes. Is this my life now?"* (Nov 25, 2025), https://www.glassdoor.ie/Community/advertising/i-just-got-feedback-from-a-client-that-is-clearly-from-chatgpt-they-uploaded-the-work-and-had-it-give-notes-is-this-my-life
- Forbes Agency Council, *20 Big Disruptors In The Agency World (And How To Adapt To Them)* (Mar 11, 2025), https://www.forbes.com/councils/forbesagencycouncil/2025/03/11/20-big-disruptors-in-the-agency-world-and-how-to-adapt-to-them/

**Evidence**
- The Glassdoor post is first-person, dated, and practitioner-written. Note what it is not: it is not a complaint about AI quality. It is a status event. The client acquired the reviewing function, and the agency noticed. In-thread responses escalate to running the revisions back through AI and to staging live anti-AI demonstrations for the client. Nobody in the thread bought a tool, so treat this as a high-confidence event and a medium-confidence purchase trigger. [E]
- In the Forbes roundup, an agency principal states the economic endgame: "You can't win a price war with someone whose costs are near zero." A second contributor supplies the judgment-versus-output distinction in agency voice: "Agencies are drowning in AI-generated insights, but collecting raw data isn't a strategy." [E]
- Corroborating trade coverage of the same widening client-side capability gap: Digiday, *Agency bosses say the AI gap with clients is only getting wider* (https://digiday.com/marketing/agency-bosses-say-the-ai-gap-with-clients-is-only-getting-wider/). Tier 2, and its subject skews larger than boutique. Use for context, not as boutique-founder speech. [E]
- The synthesis is analyst inference, and it is the load-bearing one: the defensible remaining product is the disagreement a compliant model will never volunteer to the client. It follows from the two attested quotes and is not stated by either. Do not attach a knowledge-compounding claim to it. The product has no per-client memory, so the resolution is access to other practitioners' patterns, not persistence of the founder's own. [I]

**Key phrases:** commoditization, race to the bottom, price war, near-zero costs, AI-generated feedback, judgment vs. output, pushback
- In citation or reference lines ("**Title**, Publisher, URL" and similar),
  use commas or periods as separators instead.
- In channel/criterion titles ("### N) Name, detail"), use a colon or comma.
- In prose, restructure: period and new sentence, comma, colon, or parentheses.
- Zero em dashes may remain.

Preserve everything else exactly: heading text and levels, **bold**, [E]/[I]
tags, numbers, URLs, names, quotes, bullets, ordering, and roughly the same
length. Return only the rewritten markdown.

# Sources: Barriers

## Market Adoption Frictions for AI Strategy Tools Sold to Boutique B2B SaaS Agencies (2022–2026)

Every URL below was retrieved and confirmed live on 2026-08-18. Where a verbatim quote appears in the internal dossier but could not be re-retrieved today, it is flagged as such rather than presented as verified.

### Fact 1: "We Already Have This" False Equivalence

**Summary:** The incumbent is a paid consumer assistant already open on the founder's desktop, and the failure mode is documented in adjacent-category review text: paying customers of AI marketing tools cancel because they detect no delta versus the free or already-owned model. This is the most-attested friction in the set. [E]

**Sources:**
1. Capterra, Jasper verified reviews (page current as of 2026): https://www.capterra.com/p/217242/Jasper/reviews/
2. G2, M1-Project (Elsa) reviews, the closest agency-targeted competitor: https://www.g2.com/products/m1-project/reviews

**Evidence:** A Capterra reviewer states the product's features "didn't provide any significant advantage over free alternatives like GPT or Gemini," and pairs it with a complaint about inconsistent output on a simple repeated task. [E] On the other side, M1-Project's G2 reviews show what wins this comparison: a reviewer credits Elsa with producing "a detailed Ideal Customer Profile (ICP) in under 10 minutes." [E] Note the asymmetry a seller must handle: the competitor's proof is a timed, observable artifact, while the differentiator here is a behavior (refusal to answer the question asked). Behavior must be demonstrated live in the buyer's own session, or it does not register. [I] **Verification gap:** the literal buyer phrasing "how is this different from ChatGPT?" returned only vendor-authored objection-handling content across repeated searches. Treat it as our framing, not an attested quote. [E]

**Key phrases:** "no significant advantage over free alternatives" · "I already pay for ChatGPT" · "what's it trained on" · "I'll play with it this weekend"

---

### Fact 2: "It's Fluffy and My Clients Pay Me Not to Hand Them That"

**Summary:** Rejection at this stage is aesthetic and identity-shaped, not accuracy-shaped. The buyer can recognize machine text on sight and treats indistinguishability as the defect. The fear is corroborated at scale inside the largest paid B2B marketing community. [E]

**Sources:**
1. Multiple Tabs Open (Laura Vogel, agency owner), "AI & What I'm Noticing As An Agency Owner": https://multipletabsopen.substack.com/p/ai-and-what-im-noticing-as-an-agency
2. Exit Five, "Writing for Humans: The Real AI Workflows Behind Great B2B Content": https://exitfive.com/live/writing-for-humans-the-real-ai-workflows-behind-great-b2b-content/

**Evidence:** The agency-owner post is first-person and unhedged: "I can spot AI-generated copy from a mile away. And I hate it." [E] The same post concedes heavy personal AI use, which is the whole shape of this barrier: adoption is high, trust in the output is conditional. Exit Five states on its own event page, "We surveyed 600 marketers, and nearly all share the same worry about AI: a flood of low-quality content." [E] That is the best available Tier 1 stat for this objection, and it comes from the room the buyer already pays to be in. Sales implication: grade the tool on judgment against a call the founder already knows the answer to, never on prose quality in session one. [I] **Provenance note:** the Substack quote is dossier-verbatim; today's retrieval confirmed the article and publication, not the sentence. [E]

**Key phrases:** "slop" · "starting to sound the exact same" · "weak and fluffy" · "flood of low-quality content" · anti-slop

---

### Fact 3: "It'll Just Tell Me What I Want to Hear"

**Summary:** Sycophancy is not a folk complaint. It is a documented, named model failure with a public vendor post-mortem, which is why buyers assume agreeableness by default and why a tool sold on pushback gets stress-tested in week one. [E]

**Sources:**
1. OpenAI, "Sycophancy in GPT-4o: What happened and what we're doing about it" (April 2025): https://openai.com/index/sycophancy-in-gpt-4o/
2. Georgetown Law, Tech Institute, Tech Brief: AI Sycophancy & OpenAI: https://www.law.georgetown.edu/tech-institute/research-insights/insights/tech-brief-ai-sycophancy-openai-2/
**Evidence:** OpenAI rolled back an update it described as "overly flattering or agreeable," and its follow-up post concedes that offline evaluations testing that behavior did not catch it before release. [E] Georgetown's brief records that the rollback came four days after launch, following endorsement of harmful statements. [E] Two consequences for the sale. First, the buyer's skepticism is well-founded and arguing against it damages credibility. Second, this is the one barrier where the seller can win outright by inviting the test: hand over the keyboard and ask the founder to make the tool agree with a positioning angle they know is wrong. [I] The published mechanic supports the demo, but carry the bound out loud, because no third-party verification of it exists at any tier. [E]

**Key phrases:** sycophancy · "does it ever tell me I'm wrong?" · "it just reproduces my own brief back at me" · pressure-test the positioning

---

### Fact 4: "Who Else Like Me Is Using This?"

**Summary:** Mid-cycle proof diligence is where this product has nothing to hand over. No named customer, case study, review-site listing, or quantified outcome exists at any source tier, and the product name collides with several unrelated tools that do have reviews. [E]

**Sources:**
1. G2, 2025 Buyer Behavior Report announcement (n = 1,169 B2B decision-makers, fielded April 2025): https://www.businesswire.com/news/home/20250514382531/en/G2-Report-AI-Now-Means-Always-Included-Disrupting-All-Stages-of-the-B2B-Software-Buying-Journey
2. Product Hunt: "HiveMind: Skill-based hiring on autopilot" (the recruiting product, not this one): https://www.producthunt.com/products/hivemind-ai

**Evidence:** G2's report documents "smaller deal sizes, more focused buying committees," and AI reshaping every stage of the buying journey. [E] Small committee plus small deal does not remove the proof requirement, it concentrates it on one person who has to defend the choice internally. [I] Meanwhile a founder who searches the product name lands on a different HiveMind whose page advertises reviews praising hours saved in screening. [E] That is worse than an empty result: it manufactures a false impression that the diligence was completed. Handling: state the absence, then convert it into design-partner terms with a named end date and direct access to the practitioners behind the corpus. Never borrow the parent agency's engagements as product proof, because they are human-delivered and carry zero attribution to this product. [E]

**Key phrases:** "send me a case study" · "who's your reference in B2B SaaS?" · design partner · name collision

---

### Fact 5: "I Can't Approve a Number You Won't Give Me"

**Summary:** No price is published at any tier, and the comparison set publishes entry points openly. Opacity does not stall the deal at the cold stage. It stalls it at the warmest possible moment, immediately after a good first session. [E]

**Sources:**
1. Sword and the Script: "B2B tech vendors that hide pricing may never make a prospect's short list," citing TrustRadius (April 2025): https://www.swordandthescript.com/2025/08/b2b-pricing/
2. Averi: "Introducing The AI Marketing Workspace From Averi": https://www.averi.ai/blog/introducing-the-ai-marketing-workspace-from-averi

**Evidence:** The TrustRadius finding is blunt: "the number one request from prospective buyers of B2B technology is pricing." [E] Averi publishes $45/month for its create mode plus workspace, which sets the mental category the founder will slot any tool into. [E] Absent a number, the founder defaults to the only comparable they can price: senior human strategy help, benchmarked by Go Fractional at $4,000 to $20,000 monthly retainers (https://www.gofractional.com/blog/fractional-cmo-salary), and the tool loses on risk rather than on cost. [E] Quote a firm monthly figure on the first call and put it in writing the same day, framed as tool budget and expressed per client account. Prefer a one-month paid pilot with a stated cancel path over a free trial. [I]
**Key phrases:** "what does this run per seat?" · "is there a trial?" · tool budget vs. strategy budget · shortlist elimination

---

### Fact 6: "That's Client-Confidential and You're a Vendor I Can't Vouch For"

**Summary:** The material worth bringing is client discovery and positioning work the agency holds contractually. Agency practitioners already name model ingestion of client material as the reason they limit AI use, and the governance conversation has moved from procurement to contract language. [E]

**Sources:**
1. Glassdoor Community, Advertising bowl: "Are you guys putting your decks into ChatGPT? What are you learning" (Dec 2025): https://www.glassdoor.com.au/Community/advertising/are-you-guys-putting-your-decks-into-chatgpt-what-are-you-learning
2. Digiday: "Agency AI pitches are starting to face harder questions" (June 2026): https://digiday.com/marketing/agency-ai-pitches-are-starting-to-face-harder-questions/

**Evidence:** A practitioner in the thread objects that decks "often have proprietary information about clients or agencies," naming model ingestion and public exposure as the risk. [E] The same thread contains the permission structure that cuts for the seller: this audience sanctions AI for strategy and operations work while ruling it out for creative. [E] Digiday, reporting a 3C Ventures paper, records that terms like "AI-powered" and "end-to-end" have become pitch furniture, and that data ownership, access rights and portability are moving into governance conversations. [E] Because this deal band has no procurement process, the objection has nowhere to escalate: one unanswered email about training data ends the evaluation silently. [I] Pre-empt with a one-page written answer signed by a named person, and offer the agency's own positioning as the low-exposure first use case. [I]

**Key phrases:** "where does this go?" · "do you train on it?" · AI clause in the MSA · portability · data ownership
### Fact 7: "Everything It Gives Me Stays in the Chat Window"

**Summary:** The segment's system of record is Notion or Docs, and there is no export, template output, or integration of any kind. Combined with the absence of per-client memory, the re-briefing tax quietly exceeds session value and usage decays to zero without a cancellation conversation. [E]

**Sources:**
1. Quoleady, "SaaS Content Marketing Tools We Actually Use in 2026" (agency's own stack, first person): https://www.quoleady.com/saas-content-marketing-tools/
2. Exit Five Newsletter #174, "How to build your own AI marketing assistant": https://www.exitfive.com/newsletter/how-to-build-your-own-ai-marketing-assistant-exit-five-newsletter

**Evidence:** The agency states plainly that "Notion is the center of our content operation," covering editorial calendars, briefs, client documentation, SOPs and project management, with every article starting there. [E] Any strategy output that cannot reach that surface is not in the workflow. Exit Five names the decay pattern for strategy artifacts generally: the big messaging exercise is completed, everyone aligns, "and then it just sits in a Google Doc." [E] That is in-house marketer voice rather than agency-founder voice, and should be labeled as such when reused. [E] Practical handling: build a one-paragraph saved brief block per client on the onboarding call for the founder's three largest accounts, and attach usage to a real recurring event such as Monday pitch prep. Never imply memory exists; sell the paste block as the workaround it is. [I]

**Key phrases:** "can I get this into Notion?" · re-brief · "haven't had a chance" · sits in a Google Doc · lost to non-adoption

---

### Fact 8: "This Was Built for Token Launches, Not My SaaS Clients"

**Summary:** Thirty seconds of unaccompanied diligence surfaces crypto-native positioning and a documented web3-scoped corpus. That directly undercuts the one claim the product rests on, which is that the advice comes from practitioners who have shipped this kind of work. [E]

**Sources:**
1. Hivemind (Crypto Marketing Intelligence surface): https://hivemindlives.com/
2. VoltAgent, awesome-openclaw-skills registry, `myosin-hivemind` entry: https://github.com/VoltAgent/awesome-openclaw-skills/blob/main/categories/web-and-frontend-development.md

**Evidence:** The public hero use case is nailing positioning, narrative and growth loops "before the token goes live." [E] The skills registry describes the tool as searching a "curated Web3 marketing knowledge base (RAG)" for practitioner insights, frameworks and playbooks. [E] Documented trigger conditions in the skill spec include token launch tactics and community growth. [E] Cross-industry applicability is asserted publicly only inside an FAQ answering whether the product is crypto-only, which is an objection response rather than a positioning. [E] Handling: name the heritage in the first two minutes and reframe it as adjacency. Pre-PMF launches, narrative-led categories and community-driven growth describe most Series A-B software launches too. Then stop arguing and run the founder's hardest live client problem. If the output returns crypto-shaped examples, log it and price the deal accordingly. [I]

**Key phrases:** "is this a crypto thing?" · "do you have anything in B2B SaaS?" · story-market fit · adjacency, not mismatch

---

### Fact 9: "If My Team Sees a Tool Promising Strategy Without the Agency Markup, What Am I Telling Them?"

**Summary:** The role threat at rollout is manufactured by the seller's own copy. The published anchor argues against the exact expertise the founder's senior people bill for, which is why single-seat accounts fail to expand even when the founder personally likes the product. [E]

**Sources:**
1. Hivemind, Marketing Strategy Copilot (product site, current copy): https://myosin.xyz/hivemind
2. Haus Advisors, "Agency Positioning: The Strategy Decisions a Rebrand Can't Fix" (July 2026): https://www.hausadvisors.com/blog/agency-positioning

**Evidence:** The site's own value anchor reads: "Stop paying $300/hr for advice you could get in 30 seconds," extended with "without the agency markup." [E] Said inside an agency, that is a redundancy argument. Haus Advisors describes the founder-side wound the tool should be aimed at instead, in a case narrative of a founder post-rebrand: "Same inability to explain why his agency was different from cheaper alternatives." [E] Source caveat: Haus Advisors is a rival positioning consultancy, so the narrative is competitor-authored marketing, not neutral research. [E] Context worth carrying: at a 13% average after-tax margin on average revenue of $4.43M across 31 employees (Promethean Research 2026, n = 119, sourced in the Segment appendix), senior unbillable hours are the expense the founder feels most directly, which is why the pitch must land as added senior capacity rather than substitution. [E] Seat the senior strategist first at rollout, and never repeat the "$300/hr" or "agency markup" line in front of anyone at the agency. [I]

**Key phrases:** "let me use it myself first" · "is it ethical or just efficient?" · the unfunded strategic layer · non-billable hours · one-seat renewal risk