White paper
A framework for P&C carriers adopting AI in regulated claims work — how to deploy AI on the work that affects regulated outcomes without giving up the evidence trail.
A P&C carrier in 2026 has two facts to reconcile. Regulators have moved from observation to active examination — the NAIC AI Model Bulletin is adopted in 24 states plus DC, and the NAIC AI Systems Evaluation Tool gives examiners a structured way to inspect carrier AI governance during market conduct reviews. At the same time, the claims work AI can usefully do — coverage interpretation, governed denial drafting, fair-claims-handling review, fraud screening — is exactly the work where decisions must withstand outside scrutiny.
The tools marketed to claims operations today were designed before the examination infrastructure existed. They generate decisions; they do not generate the defensible record of how each decision was made. This paper proposes the audit-bearing AI standard: which claims work AI should do, and how the evidence trail must be structured to satisfy the examination regime forming now.
The questions a carrier should be able to answer before adopting an audit-bearing AI tool — and the questions an examiner will ask.