M&A due diligence involves hundreds of contracts. The AI that reads them has a measurable cost per document. P402 makes that cost visible, governed, and settled onchain — per contract, per clause, per model tier.
This is the demo where you see the router actually picking different models for different documents and saving money. Each document card shows: type, complexity score, model assigned, cost, and why.
A senior partner sets a $50 cap for a tier-3 contract review. The cap is enforced. If projected cost exceeds the cap, the analysis is blocked before the model is invoked. The decision is logged.
Law firms billing clients per matter need defensible per-matter unit economics. This demo produces exactly that: a cost receipt for AI-assisted diligence that can be passed through to the client.
ABA Formal Opinion 512 (2024) requires lawyers using AI to understand and supervise its work. The P402 ledger provides the documentation: what model, what input, what output, at what cost, with human review recorded.
All documents are synthetic, de-identified, and clearly marked SAMPLE. No real entity names.
The infrastructure is built. The routing logic, ledger, and Tempo settlement are identical to the live healthcare demo. The legal-specific components — document classifier, clause extractor, conflict detector — are next.