Your business insurance renewal notice probably doesn't say "generative AI" anywhere on it. That's the problem. Two new industry-standard exclusion forms took effect in January 2026, and they quietly carved AI-related claims out of the general liability policies that small businesses have relied on for decades — while 74% of small businesses are already using AI tools in some capacity, often without knowing their coverage just changed underneath them.
If your business uses a chatbot to answer customer questions, an AI tool to draft marketing copy, or an automated assistant to give pricing or technical advice, you may be running that exposure completely uninsured right now. Here's what changed, why it matters, and what to actually do about it.
What Changed in Your Policy
For years, "silent AI" coverage was the default: standard commercial general liability (CGL) policies didn't mention artificial intelligence one way or the other, so claims involving AI were generally handled like any other liability claim. That ambiguity is gone.
In January 2026, the Insurance Services Office (ISO) — the organization that drafts standardized policy language used across the insurance industry — introduced two new endorsements:
- CG 40 47, which removes coverage for bodily injury and property damage claims tied to AI
- CG 40 48, which excludes personal and advertising injury claims involving AI — the category that covers defamation, copyright infringement, and misleading advertising, which is where most real-world AI incidents actually land
Carriers are adding these endorsements to policies at renewal, often without much fanfare. Between these two exclusions, four major claim categories are now potentially uncovered: bodily injury, property damage, personal/advertising injury, and products/completed operations claims that involve an AI system in any way.
Why "Customer-Facing AI" Is the Real Exposure
The risk isn't abstract. In February 2024, a Canadian tribunal ruled that Air Canada was liable after its website chatbot gave a customer incorrect information about bereavement fare refunds — the airline argued the chatbot was "a separate legal entity responsible for its own actions," and the tribunal rejected that argument outright. The company was on the hook for what its AI told a customer, full stop. That case has become the reference point for a simple legal principle: you cannot outsource liability to your AI vendor. If your chatbot makes a promise, quotes a price, or gives advice, the law treats it as if your business said it.
That's a serious problem when the tool making those promises is running with no direct human review of individual outputs. Common exposure points for small businesses include:
- A support chatbot promising a refund, discount, or service level your policy doesn't actually offer
- AI-drafted marketing copy that infringes someone else's copyright or makes a claim you can't substantiate
- An AI assistant giving technical or product advice that turns out to be wrong and someone relies on it
- A scheduling or intake bot that mishandles or exposes a customer's personal information
Insurers underwriting affirmative AI coverage are now explicitly asking whether outputs are reviewed by a human before they reach the public — marketing copy, customer messages, and technical advice are all treated as higher risk when nobody checks them first.
Your Two Options for Getting Covered
If your renewal now carries one of the new exclusions (or your broker confirms your carrier applies them), you have two realistic paths back to coverage:
1. A write-back endorsement. Some carriers will sell coverage back onto your existing CGL policy for an added premium, once they understand how you use AI and what guardrails you have. This is often the cheaper, faster option if your AI use is limited — say, an internal drafting tool with human review before anything goes external.
2. A standalone AI liability policy. For businesses with more direct AI exposure — especially customer-facing chatbots or AI-generated content going out under your name — a dedicated policy is increasingly the more direct route. HSB (a Munich Re company) launched what's described as the first major standalone AI liability product built specifically for small and mid-sized businesses in March 2026, and more carriers are expected to follow with comparable products through 2026. Coverage under these policies typically addresses AI hallucination-driven errors, IP infringement and defamation from AI output, unauthorized disclosure of confidential data, and — where relevant — bodily injury or property damage tied to reliance on AI-generated guidance.
Before you call a broker, get ahead of the underwriting conversation. Insurers are asking a fairly consistent set of questions:
- Which AI tools or systems does your business actually use, and for what?
- Are any outputs published or sent to customers without a human reviewing them first?
- What safeguards catch a harmful, false, or off-policy output before it reaches someone?
- Has your business had any AI-related customer complaints so far?
- What documentation do you have of your AI usage and review process?
Businesses that can answer these clearly — even informally, in a one-page internal policy — tend to get better pricing and broader terms than those who show up with no governance story at all.
A Straightforward Way to Reduce the Underlying Risk
Insurance is the backstop, not the fix. The businesses getting the best terms (and avoiding claims in the first place) are the ones that can point to consistent human review before AI-generated content or promises go external, and that keep a clear record of what was said, when, and by what system.
That record-keeping habit — knowing exactly what commitments your business has made and to whom — is the same discipline that keeps your books defensible when an insurer, auditor, or customer disputes a claim. If a chatbot promises a refund you didn't authorize, you need to be able to trace that transaction back to a decision, a policy, and a paper trail just as clearly as you'd trace a disputed invoice. Businesses that already track their finances with precision tend to extend that same rigor to AI governance almost by habit — it's the same instinct, applied to a newer risk.
Simplify Your Financial Management
As AI tools become part of how you run customer interactions, keeping precise, auditable records of your business's finances matters more than ever — especially if a claim or dispute ever traces back to something an automated system said or did. Beancount.io offers plain-text accounting that gives you complete transparency and control over your financial data — no black boxes, no vendor lock-in. Get started for free and see why developers and finance professionals are switching to plain-text accounting.