Malpractice-carrier renewal questionnaire — the AI-governance section

Not a statute, not a bar rule — but the item that most directly touches the firm's premium and coverage. Aon, Marsh, Beazley, and CNA legal-liability desks all added AI-governance questions to renewal questionnaires in 2025-2026. The item is short. The follow-up questions on an unsatisfactory answer are not.

What the questionnaire actually asks (representative item)

What this means in plain English

Three things the carrier wants on the record:

1. Inventory. Which AI tools does the firm run in the client-facing stack. "We don't use AI" is increasingly hard to say honestly given intake stacks, e-discovery, and drafting tools.
2. Supervision program. Who watches the AI, on what cadence, with what escalation criteria. The carrier is scoring the operational reality, not the policy document.
3. Adverse-impact audit. Evidence that the firm actively checks for segment-differential outcomes — the specific failure mode most likely to produce a claim or a class-adjacent exposure.

An unsatisfactory answer does not usually trigger denial. It triggers a supplemental questionnaire, a higher premium quote, or an AI-carve-out exclusion in the renewed policy.

What triggers the exposure in the sample

Once the C3 pattern is known (via an unrelated event, a client complaint, or a competitor's Snapshot), the carrier's post-hoc question is "when did the firm first know, and what did the firm do." The dated Snapshot answers both.

What the $499 Snapshot shows against this item

• Named AI system + audit period + intake volume — the inventory answer
• Independent-verifier evidence of per-segment audit — the supervision answer
• Signed determination + 3 fix-first items + retention plan — the adverse-impact-audit answer
• Dated declaration — the "when did the firm first know" answer

See how the Category-C determination is derived →