Across the UK, accountants say more clients now arrive with answers copied from ChatGPT style tools. These answers often appear polished and confident, which gives business owners comfort when they face tax rules or bookkeeping tasks late at night or during busy weeks.
Research published by Dext shows that this behaviour has grown quickly during 2025. 77% of accountants and bookkeepers say clients now turn to public AI tools for financial, tax or bookkeeping help. Many say clients use these tools before speaking to a professional, not after.
That change has altered conversations inside accounting firms. 72% of accountants report more clients using AI generated text to question professional advice. 68% say clients have gone further and suggested that AI could replace accountants altogether. Practitioners say these discussions often happen before any numbers are checked.
Accountants say the tone of AI answers makes the wording sounds confident even when the advice is wrong. That confidence can persuade business owners to act quickly without checking the details.
Where Do The Financial Losses Come From?
The same research shows that there definitely is financial damage. Dext found that 50% of UK accountants and bookkeepers know businesses that have lost money after following incorrect AI advice. These losses cover penalties, fines, missed allowances and incorrect payments.
Mistakes appear across everyday tasks as 46% of accountants say they regularly see errors linked to business expenses. VAT problems are a close second at 41%. Personal tax planning errors appear in 35% of cases. Payroll and business tax planning mistakes both sit at 34%.
31% of the accountants say they now see AI related errors every week. 7% deal with them daily. Only 5% say they have never come across these mistakes. Errors often look reasonable at first glance, which delays correction.
Accountants say businesses often realise something has gone wrong only after HMRC letters arrive or payments do not match expectations. At that point, correcting the problem costs more time and money.
“The damage is no longer hypothetical,” said Paul Lodder, VP accounting product strategy at Dext. “Businesses are already losing money, and accountants are spending valuable time correcting avoidable mistakes, from VAT and payroll errors to misinterpretation of expenses.
“AI has a powerful role to play in finance but there’s a fundamental difference between specialist tools built for accounting and bookkeeping, and general-purpose chatbots that don’t know a business’s true financial context.”
How Much Work Is Created Behind The Scenes?
Outside of lost money, the research shows that there is more of a drain on time. According to Dext, 93% of accountants who encounter AI driven mistakes spend up to 10 hours each month fixing them.
Time spent fixing errors varies with 44% saying they spend up to 3 hours each month correcting AI mistakes. 39% spend between 4 and 10 hours. This work often involves rebuilding records, checking calculations and rewriting filings.
Accountants say this work usually sits outside planned schedules. Businesses then face extra bills for corrective work they did not expect. Practitioners say these conversations can strain trust, even though the original issue came from public AI advice.
Many say the time spent correcting mistakes also delays other clients’ work. This creates pressure across firms during busy tax periods.
What Worries Accountants About 2026?
For the next year to come, accountants expect sharper consequences if businesses keep relying on public AI without checks. A third of those surveyed say this behaviour could trigger business failure in 2026.
Other risks stand out in the data…
43% of the experts expect more misuse of AI text to justify improper or fraudulent claims. 38% expect more fines and penalties. 37% expect closer HMRC attention after incorrect or late filings.
Also, 92% of accountants believe public AI tools should face regulation or restrictions when dealing with tax or financial guidance. 70% support formal regulation.
Accountants say the issue is not technology itself. Its more so the matter of trusting generic AI outputs for decisions that carry legal and financial consequences, without responsibility attached to the advice.
Lodder added: “If we head into 2026 with more businesses treating AI outputs as trusted tax and financial advice, without professional oversight, the consequences could be severe. The focus now should be on responsible guardrails, clearer restrictions around financial advice, and better education for businesses on what these tools can and cannot safely be used for.”
