Decision
Sell the company brain install as a productized service ladder under Hard Calls, JOLT shape but with its own pricing:
- Self-serve: guided brain install inside Hard Calls Pro ($197 one-time, already locked in the Pro composition decision). Complete instructions, nothing withheld. Higher tiers buy labor and judgment, not secrets. No second $197 SKU ever; brain content grows as modules inside Pro.
- Install: $12,500 base package plus a priced add-on rate card. Typical invoice expected to land $12,500-18,000. Page copy: “installs from $12,500”.
- Retainer: $2,495/mo, optional, requires the install (or a paid audit of a brain we did not build).
Context
The Pro composition decision (2026-08-07) already folded Hockey Stick into Pro as the brain/memory layer with “brain install w/ automations -> agency white-glove” above it, and the /free survey router already sends high-fit respondents to a booking page. This decision specifies those rungs: scope, price structure, and acceptance criteria.
Key sizing insight: install cost does not scale with headcount (the AI does the reading; 40 people’s Slack costs about what 10 people’s does). It scales with number of connected tools, number of entities, depth of history, and stack weirdness. So price on those, not on team size.
Base install scope (spelled out in the agreement)
- Brain running on the client’s accounts; they own the repos from day one
- Two access tiers: team, and owner-only for money and payroll
- Three feed connections: meetings, email, Slack
- Last 90 days of history backfilled
- Daily briefing and weekly scoreboard live
- One team training session plus a written guide
- Two weeks of tuning after go-live
Add-ons quoted at kickoff from a rate card: extra connections (CRM, support inbox, project tracker) at roughly $1,000-1,500 each, extra entities, deeper backfill, extra access tiers. A bigger company gets more line items, not a vague bigger price.
Definition of done (signed before kickoff)
- Every feed has run clean for 14 straight days (a night with nothing worth saving still counts as a clean run; the log shows it ran)
- Backfill counts match the kickoff inventory (count what exists at kickoff, write the number down; done means that many files exist)
- The 10-question demo: owner asks 10 questions about their own trailing 90 days, the brain answers each citing the page it pulled from; limited to what the connected sources contain
- Deliverables handed over: repo ownership confirmed, guide delivered, training held, two weekly scoreboards sent
Handoff review produces a written punch list against the checklist; no written objection within 7 days means accepted. Boundary: we guarantee the pipes, not the water pressure. If the client’s team stops recording meetings, that is a habits conversation for the retainer, not an acceptance fight. Brand beat: the definition of done is signed before work starts, deciding the way the product teaches.
Options considered
- Flat $15k matching JOLT Deploy: lost. Size variance lives in connections/entities/history; visible line items beat size bands and stop small companies subsidizing big ones.
- **Retainer at Studio’s $3,995:** lost. The brain retainer is monthly synthesis plus maintenance, lighter human load than Studio's production. $2,495 sits above Lore ($1,745, one channel) and below Studio ($3,995), matching scope. Priced to attach: install plus 18 months of retainer is roughly a $57k client. Raise later if attach rates are strong.
Consequences
- Funnel needs no new wiring: /free survey high-fit already routes to the booking page; the Pro brain module ends on the honest fork.
- Retainer’s recurring core is the monthly synthesis (the step that makes the brain smarter, not just bigger); scope stays the brain and its feeds or it drifts into general AI consulting.
- Open items: who the embedded operator is; whether the $12,500 floor is published on the page or quoted at booking (Hermes leans publish, the survey already filters budget and hidden pricing reads enterprise).
- Team TLDR doc created same day (Google Doc, owned michael@ryevalley.co).