E&O and AI: How Agencies Stay Covered in 2026

Are insurance policies starting to exclude AI in 2026?

Yes — beginning in 2026, standard forms and major carriers started carving artificial intelligence out of liability policies, including some errors-and-omissions coverage.

For years, AI risk was covered “silently” — no policy mentioned it, so existing cyber, tech E&O, and general liability forms implicitly absorbed it. That era is ending. Insurers are replacing silent AI with explicit language, and coverage is fragmenting across cyber, D&O, E&O, and general liability as each line narrows its AI exposure independently (Fenwick analysis). The result is a patchwork where no single policy may fully cover an AI-driven loss.

What are the new ISO AI exclusions (CG 40 47 and CG 40 48)?

In January 2026, Verisk’s ISO released endorsements that let carriers exclude generative-AI exposures from commercial general liability — a broad form and a narrower one.

CG 40 47 is the broad version: it excludes claims tied to generative-AI output under both Coverage A (bodily injury and property damage) and Coverage B (personal and advertising injury). CG 40 48 is narrower, excluding only Coverage B. A companion form, CG 35 08, addresses products and completed operations (Gallagher summary). These are optional endorsements — carriers choose whether to attach them — but availability alone signals where the market is heading.

Which carriers are already adding AI exclusions?

Several of the largest liability insurers have filed AI-exclusion language, and state regulators have approved the large majority of those filings.

W.R. Berkley introduced what commentators call an “absolute” AI exclusion for its D&O, E&O, and fiduciary-liability products, defining artificial intelligence unusually broadly — broad enough to reach almost any machine-learning system that generates predictions, content, or decisions (National Law Review). Chubb, Travelers, and Berkshire Hathaway are among the other carriers that have filed AI-related exclusions. The direction is clear even where the specific wording differs: AI is being moved from “implicitly covered” to “explicitly addressed,” and often excluded.

What does this mean for your agency’s own use of AI?

If your agency uses AI for quoting, client communication, or marketing, an AI-driven error could fall into a coverage gap the new exclusions create.

There are two exposures here, not one. The first is advisory: your commercial clients are seeing these endorsements on their renewals, and they will ask you what changed. The second is your own: agencies are adopting AI for intake, follow-up, and content, and the same exclusion trend that touches your clients touches you. An AI system that drafts a client communication or misclassifies a submission is a new source of potential error — and the coverage that would have quietly responded a year ago may now be carved out. This is general education, not legal or coverage advice; your policy language and state filings control.

How does documented AI governance help you stay covered?

It does not guarantee coverage — but it lets you show reasonable care, and a documented human-in-the-loop process is evidence your AI use was controlled, not reckless.

Underwriters, regulators, and clients are all converging on the same expectation: if you use AI, show how you govern it. That means a human reviews and approves AI-assisted work before it goes out, the scope of what AI is allowed to touch is defined, and there is an audit trail of who approved what. None of that overrides an exclusion. What it does is put you in the defensible position — the agency that can show its AI use was deliberate and reviewed, rather than the one explaining an unmanaged system after a claim. The same discipline shows up in SOC 2 AI governance, where the human approval gate, not the tool, is the control.

How Digital Monestary helps agencies govern their AI use

We build the governance layer — human review gates, audit trails, and clear scope for what AI is allowed to do — that keeps your agency’s AI use defensible as coverage tightens.

This is where an engineer-led guide matters more than a tool vendor. Our founder designed SOC 2 AI governance with human-in-the-loop review inside a regulated environment, and that same discipline is what an agency needs when it adopts AI for quoting, follow-up, and client service. The point is not to use less AI — it is to use it in a way you can stand behind.

For an agency adopting AI without a technical executive to own the oversight, that governance sits inside the Fractional AI CTO engagement, from $10,000 per month — one rung, not the headline, and most agencies start smaller. If you want to see where your agency’s AI use is exposed today, book a free demo and we’ll map it with you.