| name | deal-analyzer |
| description | Deep investment evaluation — 12-section VC analysis from pitch decks with web research, parallel AI analysis, and multi-channel delivery. |
| version | 2.0.0 |
| author | GroundUp VC |
| actions | ["analyze","evaluate","log","demo","end-demo","test"] |
Deal Analyzer v2
AI-powered investment evaluation that transforms a pitch deck into a comprehensive 12-section VC analysis. Combines deck extraction, web research, and Claude analysis to deliver investment-grade due diligence in 3-5 minutes.
Conversational Flow
Via WhatsApp:
- User sends a deck URL →
analyze runs quick extraction, sends summary, asks "Want the full 12-section investment report?"
- User replies "full report" →
full-report runs deep analysis from saved state, sends TL;DR + memo via WhatsApp + email, asks "Want me to log this to HubSpot?"
- User replies "log to hubspot" →
log searches HubSpot for the company and adds the analysis as a note
Via Email (automatic):
- Team member emails a deck to the assistant's email → email pipeline extracts deck, creates company + deal in HubSpot
- Sender receives WhatsApp with quick analysis + "Want the full 12-section investment report?"
- Sender replies "full report" → same deep analysis flow as above
When to Use This Skill
WhatsApp triggers:
- "analyze this deck: [URL]" → quick extraction
- "evaluate this deal: [URL]" → full 12-section analysis
- "review this pitch deck: [URL]"
- "run due diligence on [URL]"
- "full report" → runs full 12-section analysis from last extraction
- "log to hubspot" → logs last evaluation to HubSpot
Actions
analyze (quick — 5-15 seconds)
Extract key data from a pitch deck. Sends WhatsApp summary and asks about full report.
deal-analyzer analyze <deck-url> [phone] [sender-email]
evaluate (deep — 3-5 minutes)
Full 12-section investment analysis with web research. Delivers TL;DR + WhatsApp summary + email report. Asks about HubSpot logging.
deal-analyzer evaluate <deck-url> <phone> [email]
full-report (deep — 3-5 minutes)
Run full evaluation using saved state from a prior analyze or email extraction. No URL needed.
deal-analyzer full-report [phone] [email]
log
Log the last analysis to HubSpot as a company note.
deal-analyzer log [phone]
demo (LP meeting — one command)
Activate demo mode with two pre-cached deals: Konko AI (STRONG INVEST) and Noovox (PASS). Run once from console, then the entire flow happens via WhatsApp.
deal-analyzer demo <phone> [email]
After activating, the demo has two phases:
Phase 1 — Konko AI (STRONG INVEST):
- User sends a deck URL →
analyze uses cached Konko AI data, creates HubSpot records, sends quick summary
- User replies "full report" → progress + TL;DR + Google Doc link + auto-logs to HubSpot
- After completion, demo automatically advances to Phase 2
Phase 2 — Noovox (PASS):
4. User sends another deck URL (e.g. "David sent me this deck...") → analyze uses cached Noovox data, creates HubSpot records, sends quick summary
5. User replies "full report" → progress + TL;DR (PASS recommendation) + Google Doc link + auto-logs to HubSpot
6. User replies "end demo" → demo mode deactivated
IMPORTANT: When output starts with "DELIVERED:", the skill has already sent messages directly to WhatsApp. Do NOT send any additional messages or summaries — just acknowledge silently or say nothing.
end-demo
Deactivate demo mode. Returns to normal deal analysis behavior.
deal-analyzer end-demo
WhatsApp triggers:
- "end demo", "stop demo", "demo over" →
end-demo
test
Run test evaluation with sample deck data (FleetPulse).
deal-analyzer test
12 Analysis Sections
- Market Sizing & TAM Analysis — TAM/SAM/SOM, growth rates, market segments
- Competitive Landscape — Direct/indirect competitors, moats, white space
- Founder Background Check — Team experience, track record, red flags
- Unit Economics Deep Dive — CAC, LTV, margins, burn multiple
- Product-Market Fit Assessment — Retention, NPS, organic growth signals
- Traction & Growth Metrics — Revenue growth, user metrics, sales efficiency
- Financial Model Review — Projections, burn rate, scenario modeling
- Technology & IP Assessment — Tech stack, defensibility, data moats
- GTM Strategy Evaluation — Channels, sales model, pricing, efficiency
- Market Timing & Trends — Why now, catalysts, regulatory tailwinds
- Exit Scenario & Return Analysis — Acquirers, comparable exits, return math
- Investment Memo Summary — Recommendation, thesis, risks, next steps
Pipeline
Phase 1: EXTRACT (Claude Haiku, ~10s) — Structured data from deck
Phase 2: RESEARCH (Brave Search, ~10s) — 8 market/competitor queries
Phase 3: ANALYZE (Claude Sonnet x12, ~3m) — Sequential section analysis
Phase 4: TL;DR (Claude Haiku, ~5s) — Executive summary paragraph
Phase 5: DELIVER (~10s) — WhatsApp + email + HubSpot prompt
Delivery
- WhatsApp: TL;DR + executive summary + investment recommendation
- Email: Full ~15-page investment memo with TL;DR and all 12 sections
- HubSpot: Company note with TL;DR + investment memo (on user confirmation)
IMPORTANT: This skill sends WhatsApp messages directly. When the output starts with "DELIVERED:", all messages have already been sent. Do NOT send duplicate messages, summaries, or paraphrased versions. Just acknowledge briefly or say nothing.
Configuration
Uses shared config.yaml and .env. Required environment variables:
ANTHROPIC_API_KEY — Claude AI for extraction and analysisBRAVE_SEARCH_API_KEY — Web research for market data and competitor intelMATON_API_KEY — HubSpot integration for logging evaluations
