VERNA · Greenfield Projects

Phase completion by project

% of P0–P5 cleared

Priority builds

named greenfield horses

AutoVox LIVE · buggy Priority #1 · score 6.85

SMB sell-side revenue engine + dealership voice BDC · Vernal LLC

auto.vernallc.com ↗

SpecSetupBuildTestLiveRev

Built: ElevenLabs ConvAI (4 agents), CRM + Google SSO, acquisition engine, G1–G5 orchestration, Stripe→ledger wired
Stack: VPS Caddy/pm2/Docker (pg/redis/n8n), 4 MCP servers; repo "Car Dealership - Stack" — 1,818 files / 93 commits; leads in final_leads.csv
Deployed: auto.vernallc.com (302), crm / portal (200), ai-crm online

Blockers: ai-crm Prisma source_message_id unique-constraint bug crash-looping (164 restarts); VPS ~508 MB free RAM (OOM risk); dials hit IVR/voicemail; AUTO_BLITZ_LIMIT=0; L3 autonomous close + mobile HITL loop not built
Revenue: $0 · milestone: first paid pilot
Next: fix Prisma + RAM → deploy Stripe→ledger → caller-rep retest → 1 dealership demo → 1 paid pilot

Competitive landscape — crowded & well-funded

Numa · 1,200+ dealers, pay-for-performance voiceImpel AI · 8,000+ clients, $126M+ fundedPodium · ~$399+/moTomaMia Labs · "AI BDC employee"STELLA · Fullpath · Gubagoo

Read: incumbents own breadth and funding. AutoVox cannot win on feature breadth — its only viable wedge is a tight vertical + pay-for-performance price, and it must first clear its own stability bugs before a dealer demo.

Enterns LIVE · pre-loop Priority #2 · score 6.80

Engineering-internship marketplace (consumer/student) · enterprise tier = Vlab

enterns.com ↗

SpecSetupBuildTestLiveRev

Model: Semester 1→8, each = cert + internship scaling startup→enterprise; competency → matching → upskill, gamified
Built: Backend complete, 19/19 smoke pass: auth/RBAC, 10-Q competency engine, matching, quests/peer-review, squads, gamification, billing/Stripe, employer insights
Stack: Supabase ×11 on VPS, enterns-api (pm2), Lovable frontend; repos enterns-platform (GH) + enterns-frontend (168 commits)

Deployed: enterns.com / enterprise.enterns.com (200)
Blockers: Frontend not yet repositioned to the engineering/semester model; two-sided flow unvalidated with real users; no signed employer inventory (cold-start)
Revenue: $0 · milestone: live two-sided loop + 1 paying employer
Next: sign first employer + onboard first student cohort (5) → live loop → YC narrative

Competitive landscape — established incumbents

Handshake · 8M+ jobs, ~$5K/yr, AI matchingParker Dewey · micro-internships, 5,900+ schoolsSymplicity · $20K–$100K+ career-center SaaS

Read: incumbents own distribution and breadth. Enterns' defensible wedge is the engineering-specific, semester/cert progression + employer-guarantee model — credible differentiation, but the cold-start (supply of employer slots) is the make-or-break.

Airmo SHELL live Job-Engine track · score 5.70

Multi-tenant job-search SaaS — productization of the personal job engine

airmo.enterns.com ↗

SpecSetupBuildTestLiveRev

Pipeline: scrape → match → tailor → outreach → auto-apply, multi-tenant
Stack: Turborepo, multi-tenant Postgres (user_id everywhere), Fastify + Next, Docker; repo "airmo" — 71 files
Status: P2 Phase-1 foundation; per-user workers / OAuth vault not built

Gate: Self-imposed: fix the 0-interview yield BEFORE charging; outreach stays draft-only / consent-gated
Revenue: $0 · milestone: offer + multi-tenant ready (Clerk + Stripe)
Verdict: Doc recommendation: one will-you-pay test, else fold as a student cross-sell under Enterns

Competitive landscape — crowded, low willingness-to-pay

LazyApply · volume, no tailoringSimplify · autofill (manual submit)Jobright · autopilot + matchingAIApply · JobCopilot · Massive

Read: the 2026 market trend is "tailoring beats volume" — mass auto-apply yields mass rejection. That validates Airmo's own gate (fix interview yield first), but it's the weakest standalone bet here; strongest as a feature inside Enterns.

Wider portfolio

moonshot · enterprise · DTC

Memoria

Cognitive core / agent router · YC moonshot

Pre-deploy

Built (on Gemini): JWT multi-tenant, async SQLAlchemy models, osint.py, storage.py (S3), main.py OODA router + World Model safety gate, pgvector RAG. Blocked: local PostgreSQL [Errno 61]; alembic not run; no frontend/cloud/deploy. Score 6.00 (highest "world-changing").

Next 4 tickets: boot DB → frontend → cloud (PG+pgvector, S3) → deploy.

Vlab

Enterns Enterprise · 7-LLM council intel

Shell live

Live shell vlab.enterns.com. Pricing Free / $99 Pro / $499 Ent. Repo verna-services (389 files); data layer (PDL/Coresignal) not licensed. Competitors: Apollo, ZoomInfo (red ocean). Score 5.95.

Recommendation: repurpose internally as the outbound engine feeding AutoVox + Enterns; productize post-raise.

Nori & Oak

Circular luxury pet DTC (Japandi)

Pre-build

Fully specced: Amazon AWD + Shopify subscription flywheel, serverless AWS; COGS $9.30–10.50, $85 AOV, 47% margin. Score 4.40.

Zero-BS flag: "$100K NOI in 90 days" is infeasible vs. physical lead times — re-baselined to first revenue ~90d, $100K NOI ~6–8 months.