Skip to content

Council Verdict — Monetization Strategy (Decision Record)

Recorded 2026-07-06. This is the decision record the rest of the business docs derive from.

How it was produced

A multi-model council was convened on the question: "What is the most credible monetization and growth path for WorkWingman that preserves the help-people-first ethos?"

Seats that answered: - Codex (gpt-5.5) — with live web research verifying LinkedIn/Workday ToS, CFAA case law, and competitor status. The anchor answer. - Gemini — strategy, product architecture, "The Common" protocol vision. - Gaming-PC council — qwen3-coder:30b, gpt-oss:20b (local open models). - Streaming-PC council — llama3.3:70b (local open model, reached via a direct MCP call bypassing the harness timeout).

Chaired by Claude (Opus 4.8), synthesizing and resolving disagreements. Full transcript: ../../../llm-council/reports/COUNCIL-LOG.md.

The verdict

Best path: free local-first core → paid personal convenience → organization seats for career-support institutions → consent-gated staffing → verification → employer/ATS/verification partnerships last. Acqui-hire is an outcome, not a strategy.

Unanimous conclusions (all seats)

  1. Free local core stays real. Monetize convenience first (encrypted sync, managed connectors, AI credits, analytics) at roughly $8–30/mo.
  2. Partnerships (LinkedIn/Workday) come last. They need distribution leverage first; LinkedIn's ToS directly bans the bot/scrape/automation surface.
  3. Acqui-hire is a possible outcome, never a plan.
  4. Don't brand as "auto-apply at scale." That is what platform-bans LazyApply/Sonara/JobRight and earns employer hostility.

The one contested decision (chair ruling)

Stage 2: institutions vs. verification vs. staffing. - gpt-oss:20b argued staffing/enterprise first (agencies have budget, tolerate rough edges). - llama3.3:70b and qwen3-coder:30b ranked verification second. - Codex, Gemini, and the chair placed helper institutions (coaches, bootcamps, universities, workforce boards, outplacement) second.

Ruling: institutions second. Reasons: (a) recruiters optimize placements, not candidate welfare — leading with staffing corrupts the ethos before trust is earned; (b) verification is FCRA/consumer-reporting legal territory and belongs late; (c) the two seats ranking verification early did not reckon with that compliance cost; (d) Huntr already proves the institutional wedge works.

The load-bearing adversarial insight

Structural churn: job seekers leave the moment they are hired. This alone breaks pure-prosumer subscription economics (12–25% monthly churn; pure prosumer needs ~80–120k active users to sustain a small team). The fix is baked into the sequence: institutional seats don't churn on one person's outcome, and repositioning to lifelong career management keeps accounts alive after the hire.

What kills companies like this (converged list)

Spam-cannon brand → platform bans · ToS/CFAA exposure from automating LinkedIn/Workday · structural churn · no moat beyond autofill · connector-maintenance overwhelm · hallucinated application claims · recruiter revenue corrupting candidate trust.

  • LinkedIn User Agreement bans bots, scrapers, browser add-ons that copy the service, bypassing access controls, account sharing, and UI overlays.
  • CFAA narrowed after Van Buren (2021); hiQ v. LinkedIn gave public-profile scraping only limited, preliminary cover — not safety.
  • Workday integration is employer/partner-approved on the ATS side, not user-side automation. Full treatment: legal-compliance-risk.md.

Comparable companies

Simplify (closest free-core proof) · Teal (B2C paywall proof, acquired Ramped) · Huntr (institutional-wedge proof) · LazyApply/Sonara/JobRight (volume cautionary set) · Common App (strategic model to borrow, not a literal operating model) · Monster+CareerBuilder (Chapter 11, June 2025 — the old job-board model dying).

Infra note (acted on)

The llm-council.exe chairman seat fails when run from inside a Claude session and discards all seat answers on chairman crash. Fix applied: the exe now writes partial answers to the report before the chairman step. See ../../../llm-council/.