StrategyDraftIn reviewApprovedHemnaathSaiBoth

Buyer sequence, paid pilot hypothesis, competitive position, defensibility, founder ownership, and stop conditions.

MarkdownH2ListUse # headings, - lists, and plain text. Every save creates a revision.

# Commercial Pilot and Moat Plan ## Value proposition EvidenceGate helps hospital and payer operations prevent avoidable pre-authorization defects, verify evidence faster, and retain a replayable explanation for every finding. ## Buyer sequence ### First pilot buyer Hospital revenue-cycle or insurance operations leader with visible query delays and authority to provide a permissioned packet sample. ### Scale buyer TPA or insurer operations leader who benefits from cleaner incoming packets, governed policy execution, and reviewer productivity. ## Pilot offer - Eight to twelve week shadow-mode pilot after data access and workflow discovery. - Five document types and two procedure families. - One hundred cases. - One insurer checklist or rule bundle. - Baseline and end-line measurement. - Weekly reviewer calibration. - No autonomous decisioning. ## Commercial model hypothesis - Fixed paid onboarding for workflow mapping, document taxonomy, integration, and rule-bundle setup. - Platform subscription for governed evidence and review workflows. - Usage component per processed case or packet. - Separate pricing for insurer-specific policy maintenance and enterprise deployment controls. Pricing remains a hypothesis until ten buyer interviews and two design-partner proposals are complete. ## Competitive position Commodity capabilities include OCR, generic extraction, vector search, summarization, and agent orchestration. EvidenceGate competes on the control and evidence layer: - Fact-level provenance. - Policy as code with insurer ownership. - Deterministic quality gates. - Version-pinned RAG. - Human correction and override history. - Replay across document, model, knowledge, and rule versions. ## Data moat With partner permission, de-identified correction patterns can improve document taxonomies, extraction evaluation sets, contradiction libraries, and rule tests. Raw protected health information is not a moat and must not be reused outside contract. ## Founder validation plan - Hemnaath owns product architecture, evidence contracts, and engineering feasibility. - Sai owns hospital workflow research, annotation operations, and pilot measurement. - Both founders own buyer interviews, partner agreements, policy governance, and weekly scope decisions. ## Stop conditions - No partner will provide a lawful representative dataset. - Query reduction cannot be measured. - The product requires autonomous approval to create value. - Rule onboarding cost cannot be contained or priced. - Reviewers cannot verify findings materially faster than the existing workflow. ## Next commercial proof Secure one signed design-partner letter that names the workflow, packet volume, data-access route, baseline metrics, pilot users, security conditions, and decision date.