British firms are bleeding cash after trusting generic AI tools like ChatGPT for financial and tax advice, according to new research from Dext, with accountants and bookkeepers warning the risks could snowball into business failures in 2026.
After polling 500 accountants and bookkeepers across the UK, the accounting software platform found that half (50%) already know of businesses that have suffered direct financial losses like overpayments, penalties, and fines, as a result of incorrect or misleading AI advice.
The data suggests that misunderstanding and misuse of public AI tools for complex financial decisions are fuelling this costly risk, but despite the dangers, most firms are leaning into these “off-the-shelf” solutions.
In the past year, 77% of accountants and bookkeepers reported a surge in clients turning to public AI tools like ChatGPT and other general-purpose LLMs for financial, tax, and bookkeeping advice, while 72% have seen their clients use AI outputs to challenge professional advice.
Moreover, 68% said that some of their clients have suggested AI could replace the need for professional accounting services, hinting at a worrying future where financial decisions are made entirely by machines that have already shown a loose grasp of the context behind the numbers.
For now, this growing reliance has been accompanied by a sharp rise in errors, with consequences showing up in client finances.
Nearly a third (31%) of accountants and bookkeepers said they encounter mistakes in clients’ accounts caused by AI-generated financial or tax advice every week, with 7% seeing these errors every day, with only 5% saying they have never encountered AI slop in clients’ books.
The most common errors include incorrect interpretation of business expenses (46%), incorrectly claiming or charging VAT (41%), flawed personal tax planning (35%), payroll errors (34%) and incorrect business tax planning advice (34%).
But beyond direct financial losses, the research also highlights the strain that clients’ determination to use AI is putting on accounting and bookkeeping professions, leading to unnecessary hours billed to businesses.
Almost all (93%) of accountants and bookkeepers say they spend up to ten hours every month correcting AI errors, including 44% who spend three hours a month, and 39% spending between four and ten hours.
“Businesses are already losing money, and accountants are spending valuable time correcting avoidable mistakes, from VAT and payroll errors to misinterpretation of expenses,” said Paul Lodder, VP for accounting product strategy at Dext.
“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.”
Looking ahead, accountants expect the risks to intensify in 2026 if businesses continue relying on public AI tools without professional oversight.
Recommended reading
- AI in Finance is “Going Nowhere Fast”, Say UK Bank Leaders
- Treasury To Assess the Potential of AI in Financial Services
- Finance Leaders Turning to AI for Risk Management
A third (33%) warn of higher risks of insolvency or business failure, while others expect increased misuse of AI to justify inappropriate or fraudulent claims (43%), lead to rising fines and penalties (38%), and greater HMRC scrutiny (37%).
Nearly half (45%) believe businesses making decisions based on false confidence in incorrect AI outputs will become more common as the technology becomes more integrated into business operations and admin functions.
With concerns mounting, accountants are calling for urgent intervention, with 92% believing public AI tools should be regulated or restricted when providing financial or tax-related advice, and 70% calling for formal regulation.
“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,” added Lodder.





