Architectural Bet Ledger for Engineering Orgs

A post-decision outcome ledger for Architecture Decision Records (ADRs) and technical RFCs at 30-150 person engineering orgs. When a team merges an ADR ('we'll stay on Postgres through 50M rows because write QPS will stay under 2k'), the product extracts the stated assumptions, invariants, and falsification triggers from the ADR text itself. 30-90 days later, the engineering lead one-clicks resolution against observable signals (migration shipped? incident from the named risk? falsification trigger fired?). AE grades each ADR's reasoning against outcome and applies the 6-pattern autopsy (cosmetic confidence, premise-conclusion severing, epistemological shielding) to the ADR text — which is rich, deliberate, and written for review. Buyer: VP Engineering / CTO who wants to know which of their org's architectural bets are holding and which assumption-classes systematically decay. No per-engineer surveillance, no PR-time friction, no sandbagging incentive — the unit is the collective decision, not the individual. Code never leaves; the ADR text is the artifact AE was built to autopsy.

ADRs are the perfect substrate for AE's autopsy engine — deliberate text, collective decisions, fast-resolving, sold to a buyer who currently has zero visibility into architectural batting average.

Still in evaluation (phase: ranked). No verdict yet.