🚀 Quick Start (2 minutes)
After installing the skill and configuring the MCP server, simply ask your AI assistant a strategic question:
"Analyze whether we should enter the German market for our B2B SaaS product"
The engine automatically triggers, asks context questions, and delivers a structured consulting-grade analysis in ~6-12 minutes.
💡 Tip: The more context you provide upfront, the better the analysis. Include revenue, team size, current markets, and specific concerns.
📊 Analysis Types
Market Sizing (TAM/SAM/SOM)
"Market sizing for AI-powered compliance tools in European banking. We're a 5-person team with a working MVP, €200K ARR, targeting mid-size banks (€1-50B AUM)."
Competitive Landscape
"Competitive analysis: our project management tool vs Asana, Monday.com, and ClickUp. We differentiate on AI automation. $50/user/month pricing."
Due Diligence
"Due diligence on acquiring TechStartup Inc: €3M revenue, 20 employees, asking €15M. They do AI customer service. We're a CRM company with 500 enterprise clients."
Investment Thesis
"Evaluate this investment: pre-revenue AI company asking €5M for 20%. Claims proprietary technology, solo founder, no patents."
Strategic Planning
"We need to pivot from B2C marketplace (€50K MRR, declining 10% monthly) to B2B SaaS. 8 engineers, 3 salespeople, 18 months runway. Options and plan?"
🎯 Advanced Features
Curator Panel (Multi-Model) 5 credits
For important decisions, use the panel mode — three different AI models analyze from strategic, critical, and operational perspectives:
"Analyze [topic] with curator_mode: panel"
Three models (DeepSeek, Gemini, Qwen) produce independent sections, then an Assembler merges without repetition. Deeper, more comprehensive output.
Audience Simulation 5 credits
Test how your target audience would react BEFORE presenting:
"Simulate how Series B investors would react to this pitch" (after getting a deliverable)
8 AI personas (risk-averse CFO, aggressive growth investor, skeptical technical due-diligence lead...) read your recommendation and give honest reactions. Sentiment score tells you if you're ready.
Deep Audit 15 credits
The most comprehensive validation available — 10 phases that transform any analysis into investment-grade:
"Audit this analysis" (after decision_result)
What you get:
✓ Fact-check against real data (Polymarket, Reddit, web)
✓ Audience simulation (all stakeholders react)
✓ Enhanced deliverable (fixed contradictions, grounded claims)
✓ Confidence Delta: "Confidence: 35% → 62% (+27 pts)"
✓ Kill Criteria: "Abandon if X happens within 30 days"
✓ Contradiction Map: where sources disagree
✓ Time-Travel: "3 months ago this would have been different because..."
✓ Drift Monitor: tracks your assumptions over time
⚠️ Best for: Pitch decks before investor meetings, board presentations, M&A decisions, market entry strategies. Not needed for quick operational questions.
💡 Best Practices
1. Always provide context
The engine asks questions if you don't provide context. But you'll get faster, better results by including:
- Your company (what you do, size, stage)
- The specific decision you're facing
- Relevant numbers (revenue, users, runway)
- Your constraints (budget, time, team)
2. Attach documents
Upload PDFs, pitch decks, financial reports. They get converted to Markdown and used as grounded context. The analysis will reference YOUR data, not generic assumptions.
3. Use the right tier
| Question Type | Recommended Tier | Time |
| "Should I do X or Y?" | Fast | ~6 min |
| Market entry, partnership eval | Normal | ~12 min |
| M&A, major investment, pivot | Deep | ~45 min |
| Investor pitch, board presentation | Normal + Deep Audit | ~25 min |
4. Run the Deep Audit on anything that goes to stakeholders
If the analysis will be seen by investors, a board, or clients — run the Deep Audit. The confidence delta and kill criteria make the difference between "AI-generated" and "consulting-grade".
5. Build a knowledge base
Upload your company docs, market research, competitive intel. The engine searches your knowledge base in every analysis. Over time, it becomes smarter about YOUR specific context.