Mentor programsLive

Mentor Intelligence

Match the right mentors to every founder team, chapter by chapter, on one platform.

Role
Founder. Product, design, front-end, back-end, data and deployment.
Team
Solo build, with chapter staff as design partners
Timeline
Static prototype to production in seven shippable steps, 2026
Stack
Next.js 16, React 19, TypeScript, Supabase (Postgres, Auth, row-level security), Tailwind, Vercel, Playwright, Vitest, Gemini and Perplexity for AI features
Year
2026
Live
with its first chapter, more onboarding
Status, not a metric
1 app
one database, independent chapters that never see each other
Architecture fact
7 / 7
migration steps shipped, each one usable on its own
Delivery fact

The problem

Entrepreneur-support chapters run mentor programs on spreadsheets, email and a generic CRM. Applicants live in one sheet, mentors in another, and the three-mentor team for each founder gets assembled from memory.

Staff needed one place to track applicants, keep the roster current, see where the roster is thin, and build the right team for each pitch. A network of 130+ chapters needed the same product to work per chapter without any data crossing between them.

Key decisions

Constraint

A working static prototype already existed and chapters wanted it live fast.

Options

Keep extending the prototype; rebuild on a full stack in one go; or port it in steps.

Decision

A seven-step migration to Next.js and Supabase, each step shippable: shell and navigation, ported pages, typed models and seed data, real persistence with RLS, server-side AI, role-based visibility, then tests and CI.

Trade-off

A slower first week. In return, auth, persistence and row-level security were in place before the first real user.

Constraint

Chapters must never see each other’s data, and each wants its own brand.

Options

One deployment per chapter, or one app with tenancy on every row.

Decision

One app, one database. Every operational row carries an org id, RLS enforces it, feature flags and theme tokens brand each chapter, and a signed-in user with no membership sees nothing until invited.

Trade-off

Every query carries tenancy and a single mistake could leak, so RLS is tested in CI rather than trusted.

Constraint

AI should help staff build teams without inventing facts or writing data on its own.

Options

Let the model act on records directly, or keep it advisory.

Decision

AI runs server-side and only for staff and admins. Pitch needs are extracted from the founder’s video or document and shown for editing. The in-app assistant acts as the signed-in staff member, and only after they confirm the listed steps.

Trade-off

More clicks per task. Every change traces to a person, which is what a program director has to be able to say.

The work

Applicant tracking pipeline board
Applicant trackingApplicant tracking moves ventures through intake stages with reasons recorded at every step, and a referral to Launching OS for founders who are not ready yet. Demo data
Rally matching with mentor sign-ups, choices and capacity
Rally matchingRally matching builds mentor teams from sign-up preferences and capacity, and catches a duplicate sign-up instead of quietly replacing the first one. Demo data
Team matching with ranked mentor recommendations
Team matchingTeam matching ranks mentors for a founding team or an open seat, with the reason each one fits and the growth evidence behind it. Demo data
Roster gaps with role balance, experience-area charts and recruitment targets
Roster gapsRoster gaps turns the mentor network into a recruiting plan: coverage by experience area, balance warnings, and a recruitment brief staff can copy. Demo data
Mentor directory with availability and recent activity
Mentor directoryThe mentor directory keeps availability, hours and recent events on every profile, so nobody is matched from memory. Demo data
Company directory showing the mentors supporting each venture
Company directoryThe company directory shows which mentors support each venture and flags any venture left without a team. Demo data
Advisory board scorecard with a composite score by area
Board scorecardsBoard scorecards turn monthly assessments into a score for each area of the program, with what needs attention sorted by how much effort it takes. Demo data
Mentor Intelligence landing page: match the right mentors to every founder team
Landing pageThe public site: one platform for independent mentor programs, each keeping its own roster, intake, matching and brand.

How it was built

  1. 01DiscoverCalls and walk-throughs with the people who would use it
  2. 02CompileTranscripts, files and constraints in one requirements brief
  3. 03SpecifyA detailed, sourced spec from Perplexity, reviewed and locked
  4. 04BuildDesigned and built with the latest AI models, in the production stack
  5. 05TestUser testing on every preview, with real users on real tasks
  6. 06IterateFindings turned into the next round of changes, worst problems first
  7. 07ShipTests, security review and release, then back to testing and iterating↻ Back to Test
How I work, step by step →

Where AI was used

Gemini watches a pitch video and writes the transcript; a summarizer turns it into needs that staff edit before anything is saved. The model never writes to the database. Coding agents work under a written constitution: no production data, preview before production, public claims need evidence.

What’s next

Company workspaces for firms that run their own mentoring, and a venture’s dated track record across Launching OS, Mentor Intelligence and funding.