Resource · Regulation

What a market-conduct exam looks like

HomeResources › Market-conduct exam

How a state insurance regulator actually reviews claims handling — and the new questions they ask once AI is in the loop.

What it is

A market-conduct examination is a state insurance department's review of how a carrier actually treats policyholders — as opposed to a financial exam, which looks at solvency. For claims, that means examiners look at whether decisions were handled fairly, on time, and consistently with the policy and the law.

What examiners look at in claims

  • Timeliness: were claims acknowledged, investigated, and paid within statutory deadlines (prompt-payment rules)?
  • Fairness & consistency: were similar claims handled consistently, under UCSPA and the state's unfair-claims rules?
  • Documentation: is the claim file complete — is the basis for each decision, especially denials, written down and supported?
  • Communications: did the carrier use required disclosure language and appeal pathways?
  • Complaints: what do consumer complaints and their dispositions show?

The new layer: AI

In states that have adopted the NAIC AI Model Bulletin, examiners now also expect to see how a carrier governs the AI it uses in claims: documentation of the systems, testing and bias monitoring, human oversight, and the ability to show the basis for each AI-assisted decision. Some states use a structured tool to probe AI governance during the exam. The practical effect: "the AI made the call" is not an answer — "here is exactly how the call was made, and who reviewed it" is.

How carriers get ahead of it

The carriers that sail through are the ones whose claim files already contain a defensible record per decision: the reasoning, the policy language relied on, the required disclosures, evidence that bias was monitored, and a named human reviewer. If that record only exists in someone's head — or not at all — an exam becomes a scramble.

Velerian's governed claims agents produce that record automatically — every decision ships with a policy citation and a full evidence trail, built to be inspected. See Trust & Evidence or how governed AI agents work.

← All resources · The NAIC AI Model Bulletin, explained