Professional indemnity cover is where a firm's approach to AI stops being abstract. At the Law Society risk and compliance conference in March, brokers and underwriters made plain that they now treat a firm's use of AI as a live underwriting question, and that how you govern it bears on the terms you are offered. For a smaller practice heading into renewal, a loose habit around these tools turns into a cost.
What the insurers said
The message from the market was blunt. One delegate poll at the conference found 14 per cent describing AI at their firm as allowed but largely unmanaged, a figure the insurers in the room called alarming. Underwriters want clear and accountable governance rather than a vague sense that people are experimenting. At renewal they are asking for detail on your AI policy, the controls that stop unauthorised use, and how staff behave with the tools in practice. A firm that cannot answer looks like a firm carrying a risk it has not measured, and that reads straight through to the premium.
Why it is landing now
None of this arrives from nowhere. The past year has produced a steady run of cases where AI-generated material, false citations and invented authorities, reached the court and drew judicial criticism. Insurers follow those cases closely, because each one is a claim waiting to happen and a sign of how a firm's controls hold up under pressure. When the loss that worries an underwriter is a solicitor filing something an unchecked tool produced, the questions on the proposal form follow the same fault line. That is why the AI section of a renewal has grown from a single tick box into a set of pointed questions, and why a firm that treats it as an afterthought stands out for the wrong reason.
The accountability gap
The same session exposed where the risk concentrates. Asked who is responsible for managing AI use, almost half of those polled pointed to the individual fee-earner, and only around a quarter to the supervising or managing partner. That is the wrong way round for a regulated firm. Responsibility that sits with the person using the tool, and not with the people who supervise the work, is the pattern an underwriter reads as weak control. It is the pattern the regulator warns against too, because a named person has to stay accountable for AI-assisted work rather than the duty dissolving across the desks.
What underwriters are weighing
Strip away the detail and insurers return to three things. They want to know that AI-produced work is checked for accuracy before it leaves the firm, that client data stays secure and does not pass into a tool that mishandles it, and that a person verifies the output rather than trusting it on sight. These are the duties the profession already carries, so the insurer is not inventing a new standard. It is asking whether you can show you meet the standard you already hold. The SRA's compliance tips on AI and technology set out the regulator's expectations in much the same terms an underwriter uses.
What a smaller firm should do
You do not need a thick manual to answer a broker well. You need a short written AI policy, a record of which tools the firm uses and for what, and a clear line showing who checks AI-assisted work before it goes out. Put those in place before the renewal form arrives, rather than in a scramble the week it does. A firm that hands its broker a clear account of how it controls AI presents as a lower risk, and lower risk is what keeps cover affordable. Treat this as the year your AI housekeeping starts to show in the premium, because on the evidence from the market it already does.
If you want your AI controls in order before your next renewal, and set out in the terms an underwriter asks about, that is the work our compliance review does: see how it fits your firm.