KPMG expands AI-based audit push in UK after FTSE 250 pilot

KPMG expands AI-based audit push in UK after FTSE 250 pilot
KPMG pushes AI audits

Audit firms are testing how artificial intelligence can reshape a labor-intensive part of the accounting market as clients and regulators weigh potential efficiency gains against oversight risks. KPMG says an AI-led pilot on the financial statements of an unnamed FTSE 250 company shows the technology can probe some issues more deeply, though human intervention remains necessary.

Highlights

  • KPMG completed its first AI-based audit of a FTSE 250 company in 2024 using over 100 autonomous AI agents through Project Exponential.
  • KPMG plans to expand AI-based audits to up to 20 larger clients but notes limitations including data quality and technology scalability for mass adoption.
  • UK regulators including the Financial Reporting Council warn AI may threaten auditor integrity, with 23 percent of audit committee chairs reporting limited confidence in the technology.

Pilot audit results and rollout plans

As first reported by Financial Times, KPMG has completed its first AI-based audit of a UK-listed company, deploying more than 100 AI agents this year to review the financial statements of an unnamed FTSE 250 business alongside a conventional audit by its accountants. Cath Burnet, KPMG's UK head of audit, says the system in several cases goes deeper than the regular audit team and asks better questions on aspects of the client's finances.

KPMG says the exercise, codenamed Project Exponential, uses AI across the full audit process rather than for isolated tasks. The agents gather data, analyse and document it, then form draft conclusions and observations, while also suggesting next steps as they work autonomously with each other.

The firm says a team builds the agents in four months and an audit partner signs off on the experimental audit, although KPMG does not rely on that version when it formally signs off the company's accounts. Burnet says the project team has a different shape and size from a traditional audit, and KPMG plans to expand the model to as many as 20 larger audits with the aim of eventually offering AI-based audits to clients.

Regulatory scrutiny and limits on adoption

Human oversight remains central to the process. Burnet says a person has to step in during one part of the pilot after an AI agent fails to match a set of regulations to company documents, while staff set the agents' objectives, provide business and prior-audit information, check their work periodically and challenge their conclusions before the audit opinion is finalised.

KPMG says there are clear time-saving opportunities from AI-based audits, but it declines to say whether fees would fall as a result. Burnet also says the technology is not yet available at the scale needed to run thousands of audits, some companies may prefer not to use the service, and few currently have data clean enough for AI agents to handle effectively.

The broader push comes as UK regulators and audit committees raise concerns about over-reliance on artificial intelligence. The Financial Reporting Council warns this week that AI could pose a distinct threat to auditor integrity and independence if firms use insights from AI agents to advise audit clients, while Mark Babington, the FRC's executive director of regulatory standards, says auditors remain accountable for how the technology is used and must be able to explain and challenge its outputs.

A recent report from the Centre for Public Interest Audit finds that 23 per cent of company audit committee chairs have limited confidence in AI. The concerns include weaker professional scepticism, data security risks and possible hallucinations, while KPMG says its system provides full visibility and traceability by recording the agents' interactions with documents down to individual characters and timestamps.

Earlier, we outlined how AI-era software changes can heighten outage risks across fragile digital infrastructure in banking and aviation.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.