FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Apex Insurance Brokers
Speak to a brokerGet a quote →

Aipedia: Accountants

Using AI chatbots for FRS 102 and technical queries

Accountants can use AI chatbots for technical research, but the answers must be checked against the standard or legislation itself. Wrong technical advice is one of the most common sources of negligence claims against accountants.

In short

Generative AI chatbots answer questions about FRS 102, IFRS and company law in confident, well-structured prose. They can also invent paragraph references, describe superseded requirements or apply the wrong regime. The PCRT AI guidance (January 2026) says to treat AI output as a junior colleague's draft, and the CCAB draft Statement says never to accept output at face value and to check any sources the AI cites. Elden v HMRC shows the consequences: AI-produced case summaries were materially wrong and the Tribunal called it professional incompetence. Wrong technical advice is a classic PI claim.

How accountants use AI for technical research

A general-purpose generative AI chatbot is often the quickest way to get a first answer to a technical question: how to account for a lease under FRS 102, whether a company qualifies for small company exemptions, how a group reconstruction should be presented, or what a change in company law means for a filing. Some practice tools add a chatbot trained on standards and commentary.

The answers read well, but chatbots predict likely text rather than look up the rule. That creates two problems. The first is hallucination: invented paragraph numbers, made-up cases or misquoted thresholds. The second is currency: the tool may have learned from material written before an amendment took effect and present the old position as current.

The duties that apply

What can go wrong

The leading example involves tax rather than accounting standards. In Elden v HMRC [2026] UKFTT 41 (TC) a firm of ICAEW-regulated chartered accountants used AI in a skeleton argument, and the case summaries it produced were materially wrong. The Tribunal said at para 139 that this "lack of sufficient care amounts to professional incompetence on the part of any regulated individual or firm involved".

Illustrative scenario: a manager asks a chatbot whether a client's arrangement qualifies for a disclosure exemption. The answer cites a paragraph that does not say what the chatbot claims. The firm relies on it, the client omits a disclosure, and the omission later causes problems with a lender and a costly restatement. The firm has no record of what it asked or what it checked.

How PI usually responds

Wrong technical advice is core PI territory, and a civil liability wording usually responds whether the error came from a textbook, a colleague or a chatbot, subject to the terms. Insurers and defence lawyers will want to see what research was done. A prompt history and a note of the primary sources checked can make a claim easier to defend or settle. See AI hallucination liability and what to document about AI use. Typical claims are described in accountant negligence claims.

Practical controls

How Apex helps

Apex arranges PI for accountants, and a broker can explain how your policy treats advice informed by AI research. See the accountancy AI overview and the AI guidance explainer.

Sources

Frequently asked

Is it negligent to use an AI chatbot for technical research?

Not in itself. The question is whether the conclusion was checked against primary sources before it was relied on. The Code requires competence and a judgement on whether technology output is appropriate for its purpose, and Elden shows tribunals expect that care.

Why do chatbots give out-of-date answers on standards and law?

They generate text from what they learned during training, which may predate amendments to standards or legislation, and they do not always say so. Always check the effective date of any requirement they describe against the current text.

Should we keep screenshots of prompts?

The PCRT AI guidance suggests keeping an audit trail of AI tools used and screenshots of prompts. For significant technical points this record shows what you asked, what you were told and what you checked, which helps if the advice is later challenged.

Would PI cover a claim based on a hallucinated reference?

Usually a civil liability policy responds to negligent advice whatever its source, subject to the terms and any exclusions. Being able to show the research process and the checks you made can help your insurer defend or settle the claim.

Check your PI covers how you use AI

Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.

Start a PI quote Request a callback

Apex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.