• [firm] Verdict
  PASS · 22/100

  Three concurrent, independently-fatal problems converge on a pass. (1) Market is occupied: Zermit AI launched March 2026 into the identical SMB+homeowner AI permit-prep niche, is live in 9 cities including two of Permitly's planned launches, and runs a contractor marketplace flywheel — Permitly enters as an undercapitalized second-mover with no articulated differentiation. (2) Core infrastructure premise is wrong by an order of magnitude: production Accela write-API access requires $10–50K and 3–6 months of partner certification, not $3K in 90 days. (3) The $35K seed cannot legally launch — five-state UPL opinions ($15–25K), E&O insurance, ICC code licensing, Colorado AI Act assessment, and Accela commercial agreement collectively exceed the entire round before a line of code is written. Unit economics also degrade meaningfully once jurisdiction-data maintenance ($100–250K/year at 50 cities) and infrastructure are modeled, converting stated Year-2 profit into a likely loss. A 'FirstPass' brand promise in a domain with 30–60% baseline first-submission rejection compounds the risk with FTC/UDAP exposure.

• [firm] Next step
  Next step

  Decline current $35K seed terms

• [firm] Next step
  Next step

  Send founder a detailed kill-shot memo covering Zermit competitive landscape, Accela partner reality, and compliance budget gap

• [firm] Next step
  Next step

  Offer a 12-month re-evaluation window contingent on the five monitor conditions above being met

• [firm] Next step
  Next step

  Flag Zermit AI for separate diligence as a potentially more capital-efficient entry into this thesis

• [dev] Fatal flaw
  Fatal flaw

  Zermit AI launched March 2026 — 2 months before this analysis — into the identical market (AI permit prep for residential contractors and homeowners, $99-range pricing, conversational AI interface) and is already live in 9 cities including Austin and Seattle, two of Permitly's planned launch cities. Permitly is not a first-mover entering a gap; it is a late entrant at $35K competing against an operational, geographically expanding competitor with a contractor lead-gen marketplace flywheel already generating network effects. No differentiated wedge is articulated that would overcome a 9-city head start with a 70:1 funding disadvantage.

• [dev] Fatal flaw
  Fatal flaw

  The Accela API production access assumption is wrong by an order of magnitude. Commercial submission-capable API access requires Marketplace partner certification ($10–50K, 3–6 months minimum) or per-jurisdiction OAuth negotiation with individual city IT departments — the plan budgets $3K and 90 days. This is not a cost overrun; it is a fundamental misunderstanding of Accela's partner program structure. Without production write-API access, programmatic permit submission — the core product mechanic — cannot be built. The technical premise collapses before engineering begins.

• [dev] Fatal flaw
  Fatal flaw

  $35K seed is insufficient by 4–7x for the described scope. A single-city production MVP (Accela integration, RAG pipeline, document assembly, webhook tracking, reuse library, frontend) realistically costs $150–250K over 6–9 months. Engineering at market contractor rates ($80–130/hr) exhausts the full $11K line item in 3–4 weeks. Hosting, infrastructure, and the $100–250K/year jurisdiction data maintenance burden at 50-city steady state are entirely unmodeled. The financial plan is a prototype budget dressed as a production roadmap.

Showing first 8 of 19 decisions.