Accountex 2026: the AI questions partners haven’t answered

I went to Accountex last week expecting to find the accounting profession deep in AI strategy. Debating implementation, arguing about risk, wrestling with what it means for how firms price and staff and compete.

What I actually found was a profession still working out what AI means for it. Change at this scale takes time to land, and the fact that these conversations are happening at all is progress. But there is a meaningful difference between processing the arrival of AI and what needs to be locked down before the technology goes any deeper into client work.

What is your liability position when AI is wrong?

This is the first question to answer, and most firms have not answered it. Professional indemnity policies were written before AI became part of how accountants work. Whether your insurer considers AI-assisted outputs as falling within your existing coverage is not something to assume. It is something to confirm in writing.

The risk is not theoretical. The moment a client disputes advice that was partly generated or informed by an AI tool, your policy wording matters. Get on a call with your PI insurer before you go any further with implementation.

Do your engagement letters cover AI error scenarios?

Your engagement letter is your legal protection with every client. Most were drafted for a world where the work was done by a human professional applying trained judgement. If AI is now part of that process, your letters need to reflect it.

At minimum, your letters should clarify what role AI plays in your workflow, who takes professional responsibility for the output, and what the review process looks like before work reaches the client. This is not about frightening clients. It is about being clear, which is what good professional practice requires.

Do you know where your client data is going?

Most AI tools process the data you feed them. If you are uploading client financial information, payroll records, or tax data into any AI system, you need to know where that data is stored, how long it is retained, and whether the tool’s data processing terms are compatible with your GDPR obligations.

This is not a niche concern. It is a live regulatory exposure. Your clients have not consented to their data being processed by a third-party AI system unless you have told them and documented it. Check the terms of every tool you are using and update your privacy notices accordingly.

Who signs off on AI output before it reaches a client?

AI tools produce outputs. Those outputs need a qualified professional to review them before they become advice, correspondence, or a filed return. If that review step is not formally assigned, documented, and consistently applied, you have a quality control gap.

The firms that will handle this well are the ones that treat AI as a tool requiring professional judgement to operate, not a tool that replaces it. That distinction matters for quality, for regulatory compliance, and for what happens if something goes wrong.

What I took from the day

What was on the agenda at Accountex 2026 tells you exactly where most of the profession is right now. Processing the change, managing the anxiety, building basic capability. That is a legitimate place to be. But the window for getting comfortable is shorter than it feels. The firms that come out ahead will not be the ones that waited until these questions became urgent. They will be the ones that sorted the foundations before the pressure arrived.

Lexendo

Find the risk across your client portfolio — before HMRC does.

Six domains of UK tax and compliance coverage. 198 HMRC sources. 265 tribunal decisions. 30-day money-back guarantee.

If Lexendo doesn't find a risk worth acting on in your first 30 days — full refund.