A founder tells us everything when they apply and onboard — their goals, their gaps, their capital picture, the advisor they wish they had. Today that intelligence sits in forms. The Founder OS activates it: every answer a founder gives us becomes capital matched, curriculum shaped, advisors paired, and progress tracked.
This started with the Turner call (VelocityTX, July 22): one programs coordinator, no coding background, built a founder CRM + funding matcher that runs his whole portfolio. Our version is bigger because our data is bigger — 122 companies this fall, every onboarding record joined to its full application. The principle came out of that call and hasn't changed: don't let founder information sit static on a form. Activate it. A founder should never tell us something twice, and nothing a founder tells us should fail to come back to them as help.
The matcher tests each founder against every funding source's real rules — identity eligibility, revenue floors and ceilings, business age, geography — and produces a ranked shortlist. Anything it can't verify goes to a human, never asserted. Proof run on the 2025 Spring cohort: 44 of 90 founders had an actionable best-fit grant, ~3 each. Dakota approved.
Every application answers five questions we currently read once and never look at again: 6–12 month goals · why now · problem evidence · support areas · the key question they need answered. The OS turns that into a living goal record on each founder — visible to staff, referenced at check-ins, measured at midpoint and graduation. "Did we answer their key question?" becomes a reportable number.
Founders told us their top challenges, the topics and speakers they want, and — the underused half — the skills they're willing to teach each other. Aggregated across a cohort, that's a session plan: the challenge clusters say what to program, the speaker asks say who to invite, and skills-to-share turns 122 companies into faculty for each other. Programming stops being a best guess.
Two sides, both already structured. Founders state the advisor they wish they had, in their own words. The roster answers with 190 advisors, 96% tagged across a 77-skill taxonomy, self-sorted into three seat types — subject-matter expert, industry expert, been-there entrepreneur — with real capacity history (who's been asked, who said yes, who's tapped out). The matcher assembles a cabinet per founder instead of the same twelve names getting every call.
The OS reads and writes ClickUp; it never becomes a second source of truth. Machine fields are Claude's; staff status and notes are never overwritten. If the OS breaks, the data is safe and editable where it always was.
Engagements recorded at the moment they happen rebuild "founders served," jobs, and capital numbers for the annual report and funders — with zero year-end reconstruction. The reporting is a byproduct, not a project.
44 deferred applicants are already a named Spring 2027 pipeline. Advisor pairings, goal outcomes, and curriculum asks all persist — so next fall starts from evidence, not memory.
| Piece | State | Next step |
|---|---|---|
| Fall 2026 dataset | Complete — 122 companies, apps joined, refreshable from ClickUp | Stays current via one-command refresh |
| Capital matcher | Proven on 2025 cohort; Dakota approved | Run against Fall 2026; write matches to ClickUp |
| Advisor matcher | Data validated both sides; build session with Amy | Blind-test against Amy's answer-key cabinets |
| Goals system | Data in hand (122/122) — new scope, needs design sign-off | Define the goal card + midpoint/final check-in flow |
| Curriculum view | Data in hand — new scope, needs design sign-off | Cluster fall challenges/topics into a per-program readout |
| OS interface | Working prototype (anonymized data), @ec.co gating specced | Wire to live ClickUp behind Cloudflare Access |
| Field cleanup | Checklist with Dakota | Gate for wiring the OS to live data |