The renewal application arrived, and it did not look like last year's. Tucked between the familiar questions about practice areas, claim history, and headcount was a new section: which AI tools does the firm use, who supervises them, and does the firm have a written policy for how attorneys check machine-generated work. The managing partner forwarded it to the general counsel with two words. What changed?

Plenty. Carriers that spent decades pricing legal malpractice against a fairly stable set of inputs are now pricing a moving one: the risk that a lawyer files a brief with a fake citation, leaks privileged information into a public model, or leans on a drafting tool without catching what it got wrong.

The result is a wave of underwriting questions, new endorsements, and exclusions that firm leaders have to think through before their next renewal.

Decide Whether to Disclose Every AI Tool in Use

The first decision arrives before the policy is even quoted. Renewal applications now routinely ask what AI tools attorneys use, and the honest answer at most firms is longer and messier than partners expect. Associates paste passages into consumer chatbots. Paralegals summarize deposition transcripts with a browser extension nobody vetted.

A litigation group runs a private pilot on an internal model.

Underreporting is tempting and dangerous. One industry analysis of the 2026 EPIC lawyers' professional liability claims survey found that seven of the 13 major insurers covering more than 80% of Am Law 200 firms reported a rise in AI-related claims over the past year, and nine-figure malpractice payouts are no longer freak events.

Carriers are watching. If a claim arises from a tool the firm did not disclose, the coverage fight gets ugly fast. The practical move is a shadow-IT audit before the application goes back. Ask every practice group what they use, in writing.

Disclose the full list, and be ready to describe the guardrails around each one.

Choose Between Silent Coverage and an Affirmative AI Endorsement

Most legal malpractice policies in force today do not mention AI at all. That silence has so far been read as coverage: a negligence claim tied to a hallucinated citation looks like any other negligence claim, and a Utah State Bar analysis notes that lawyers' professional liability policies typically do not currently exclude coverage for negligence claims tied to generative AI use.

Silence cuts both ways. When the form has no AI language, the carrier decides at claim time whether the loss fits an existing exclusion for intentional acts, for cyber events, or for services outside the scope of legal practice. Firms now face a real choice at renewal:

• Stay silent. Keep the existing form, accept that the carrier's position on AI claims is unwritten, and hope your factual record at claim time is clean enough to win the argument.
• Ask for an affirmative endorsement. Push the carrier to say, in the policy, that covered professional services include the use of AI tools under attorney supervision. Expect governance conditions attached.
• Accept a manuscript exclusion. Some carriers will offer renewal only with an AI carve-out, which is usually worse than it sounds. Read the wording carefully before signing.

Set a Verification Standard You Can Actually Defend

Underwriters have stopped asking whether attorneys use AI and started asking how they supervise it. A firm that can produce a written verification protocol, a training log, and a record of who reviewed what is a different risk than a firm that cannot, and pricing is starting to reflect that.

The verification rule does not need to be elaborate. It needs to be specific, followed, and documented. Every cited case checked against the actual reporter. Every AI-drafted clause reviewed by a named attorney before it leaves the building.

Every prompt that touches confidential material should be routed through an approved, contained environment rather than a public tool. Firms building this out often lean on outside guidance, and Law.co's look at AI in the legal industry look at AI in the legal industry is a useful starting point for partners trying to understand where machine work reliably helps and where human review is non-negotiable.

Build the Governance Record Before You Need It

The firms getting the best renewal terms in 2026 are not the ones using the least AI. They are the ones who can prove, on paper, that they use it deliberately. A written AI policy. Named partners accountable for tool selection.

A training record showing every attorney has been through the confidentiality and verification basics. An incident log, even if the log is mostly empty.

The economics of legal work are shifting because AI genuinely does compress document review, discovery, and first-draft production. That is not going to reverse. What will change is which firms capture the upside on favorable insurance terms and which ones pay more, or lose preferred-risk standing, because they treated governance as an afterthought.

The renewal application is where that sorting happens. Answer it like the underwriting document it now is.