KPMG finds AI agents ask better questions than human auditors in trial

Nov 18, 2026 - 13:00
0 1
KPMG finds AI agents ask better questions than human auditors in trial

KPMG has found that artificial intelligence (AI) agents asked better questions and investigated certain financial issues more deeply than human accountants during an experimental audit of a UK-listed company. The Big Four accounting firm deployed more than 100 AI agents to examine the financial statements of an unnamed FTSE 250 company this year, the Financial Times reported on Wednesday.

The experiment, conducted alongside a conventional audit by KPMG's accountants, involved using AI agents across different stages of the audit process. KPMG's UK head of audit, Cath Burnet, told the Financial Times that in some instances, the AI agents were able to investigate issues more thoroughly than the regular audit team.

"In several cases the AI was able to go so much deeper," Burnet said, adding that the technology had asked "better questions" in some situations.

However, human intervention was still required during the exercise. Burnet told the publication that an auditor had to step in when an AI agent failed to match a set of regulations with the company's documents.

How KPMG used AI agents in the audit

Codenamed Project Exponential, the initiative used AI agents across the entire audit process rather than for individual tasks. The agents collected data, analysed and documented it, and prepared draft conclusions and observations. They also worked autonomously with one another and suggested subsequent steps.

Also Read

According to the Financial Times, a team developed the agents over four months. An audit partner signed off on the experimental exercise, but KPMG did not rely on its findings when formally approving the company's accounts.

Human oversight remained part of the process, the report said. Staff set objectives for the agents, provided information about the company's business and previous audits, reviewed their work and challenged their conclusions before finalising the audit opinion, it added.

Burnet said the experimental team had a "different shape and size" compared with a traditional audit, but added that human involvement remained "critical".

KPMG plans to extend the project to as many as 20 larger audits and eventually offer AI-based audits to clients.

AI adoption raises questions over audit costs and oversight

Audit firms traditionally employ large teams of junior accountants to carry out labour-intensive work under the supervision of senior staff. KPMG acknowledged that AI-based audits offered "clear time-saving opportunities" but did not say whether these efficiencies would lead to lower audit fees.

Rival accounting firm PwC has also conducted parallel audit projects to assess AI's performance across the process. "To see how AI performs across the whole process without relying on it," Andy Hammond, PwC's global assurance leader, told the Financial Times.

However, KPMG said the technology was not yet available at the scale required to conduct thousands of audits. Burnet also said that some companies did not have sufficiently clean data for AI agents to work effectively.

The UK's Financial Reporting Council (FRC) has warned that sharing insights generated through AI and other technologies with audit clients could create risks to auditors' integrity, objectivity and independence, particularly when the work producing those insights falls outside the financial statement audit.

"Whatever technology a firm uses, the auditor remains accountable," Mark Babington, the FRC's executive director of regulatory standards, said, according to the Financial Times. He added that auditors must be able to explain how they used the technology, challenge its outputs and stand behind the evidence.

A report by the Centre for Public Interest Audit, published in October, found that 23 per cent of company audit committee chairs had "limited" confidence in the use of AI in audits of their companies' financial statements. Concerns included the risk of auditors becoming over-reliant on AI, weakening their professional scepticism, as well as data security and the possibility of AI-generated false or fabricated information.

Burnet said KPMG's system provided "full visibility and full traceability" of the AI agents' work, recording their interactions with documents down to individual characters and timestamps.

What's Your Reaction?

Like Like 0
Dislike Dislike 0
Love Love 0
Funny Funny 0
Wow Wow 0
Sad Sad 0
Angry Angry 0

Comments (0)

User