What Happens When Your PI Insurer Asks About AI Governance

The Renewal Conversation That Is Changing

For most professional services firms and technology companies, professional indemnity insurance renewal has been a largely administrative exercise. Submit the renewal form, confirm the practice areas have not changed materially, and receive the updated policy.

That process is changing.

Underwriters at major PI carriers - particularly in the UK, Australia, and the EU - are beginning to introduce AI-specific questions into renewal questionnaires. Not as standard practice yet, but as a growing trend that will be standard within 18 to 24 months.

The organisations that are unprepared for these questions face two outcomes: coverage exclusions they did not anticipate, or premium increases that reflect an underwriter's uncertainty about AI-related liability exposure.

What PI Underwriters Are Actually Worried About

Professional indemnity insurance covers claims arising from errors, omissions, and negligence in the delivery of professional services. The coverage question with AI is straightforward: if an AI system makes an error that results in client loss or harm, who is liable?

The answer depends on whether the professional exercised appropriate oversight and used the tool in a manner consistent with the standard of care expected in their profession. That determination requires evidence. And evidence requires governance.

Underwriters are trying to assess three things when they ask about AI governance:

First, whether AI is being used in ways that materially change professional liability exposure. Using AI to draft routine correspondence is a different risk profile from using AI to generate clinical recommendations, legal analysis, or financial advice that a client relies upon. Underwriters want to understand the decision-making role AI plays in the work product.

Second, whether the organisation has controls that would support a defence. If a claim arises from an AI-assisted output, can the professional demonstrate they reviewed, verified, and exercised professional judgement on that output? Documentation of review processes is the foundation of any defence.

Third, whether the organisation understands its own AI exposure. An organisation that cannot describe how its AI systems are used, what data they were trained on, and what review processes are in place is presenting as an unknown risk. Unknown risks get excluded or priced conservatively.

The Questions Being Asked at Renewal

Based on emerging practice across major carriers, the questions appearing in AI-specific renewal questionnaire sections include:

• •Do you use AI tools to assist in the delivery of professional services to clients?
• •If yes, which functions? (Options typically include: drafting, analysis, advice generation, document review, data processing, client communication, other)
• •Do you have a documented AI use policy governing how staff may use AI in client work?
• •Do you require human review and sign-off on AI-generated outputs before they are delivered to clients?
• •Have you assessed the AI tools you use against relevant regulatory frameworks?
• •Has your use of AI been subject to any client complaint, regulatory enquiry, or dispute in the past 12 months?

The challenge for most organisations is not answering yes or no. It is that they have not thought through these questions systematically and cannot provide consistent, documented answers.

What the Documentation Gap Actually Costs

A professional services firm that uses AI tools without an AI use policy, without documented review processes, and without any governance framework faces a specific risk at renewal: the underwriter may conclude that AI-related claims are either excluded from the policy or covered only at a significantly higher premium.

Exclusion language being seen in the market includes clauses that exclude coverage for claims arising from the use of AI tools where the organisation cannot demonstrate documented oversight procedures.

For a firm billing $5 million annually with a standard PI limit of $10 million, the difference between covered and excluded on an AI-related claim is the entirety of the claim. An exclusion that seemed like boilerplate at renewal becomes material when a claim is made.

The Governance Documentation That Addresses the Questions

The documentation needed to answer PI renewal questions about AI governance is not a compliance exercise distinct from your actual practice. It is a description of how you actually work - or how you need to work - with AI tools.

The minimum documentation set that addresses underwriter concerns includes:

An AI use policy that defines which tools are approved for use in client work, what functions they may be used for, what they may not be used for, and what review requirements apply before AI-assisted outputs are delivered to clients.

A review procedure document that specifies how professional review of AI-assisted outputs is conducted and recorded. This does not need to be elaborate. It needs to exist and be followed.

A risk assessment of the AI tools in use, covering the training data source, known limitations, and the categories of work they are applied to. A structured diagnostic report that maps your AI use against regulatory frameworks serves this purpose and creates a dated record.

An incident log recording any instances where AI-assisted outputs were identified as incorrect, incomplete, or potentially harmful, and what action was taken. This log demonstrates active oversight even when things go wrong.

The Timing Issue

The time to have this documentation in place is before renewal, not after a claim. An organisation that introduces AI governance documentation in response to a claim faces the additional challenge of demonstrating that the governance was genuine rather than retrospective.

If your PI renewal is within the next 12 months, the conversation with your underwriter about AI governance is already relevant. If your renewal is further out, the time advantage is available but narrowing as the market standardises its approach to AI risk.

A structured AI compliance report is not a substitute for professional legal advice about your PI coverage. It is the starting point that makes every other conversation faster - with your insurer, your clients, and your regulators.