Skip to content
NewA calmer, faster ZeroHour

KPMG audit trial finds AI agents dug deeper than staff

Original title: KPMG's "Project Exponential" put AI agents alongside a real audit team and found the bots dug deeper on individual questions

The Decoderby Maximilian Schreiner ()3 min read
Part of a story covered by 2 sourcesKPMG UK trial finds AI agents ask deeper audit questionsMerged summary & timeline
AI summary · x-ai/grok-4.7Summaries by Vani AI

Humans still set goals and signed the opinion; KPMG may extend the trial to as many as 20 larger audits.

KPMG UK ran “Project Exponential,” placing AI agents alongside a real audit team. The agents collected, analyzed and documented data and drafted conclusions; in several cases they went deeper or asked better questions than auditors, though a human intervened when an agent mismatched regulations to company documents. The firm says oversight stayed critical and scale is limited by data quality. The UK FRC warned that extra findings from AI audits can threaten independence if shared like a consulting pitch.

Unconfirmed: KPMG has not said whether time savings would reduce client audit fees.

Key facts
  • Two partners and a director worked with senior managers and junior staff.
  • KPMG says interactions are logged down to characters and timestamps.
  • The FRC says mandatory control-weakness reports are not non-audit services.
  • KPMG cites time savings but will not say whether client fees would fall.
  • PwC has run its own parallel audits, the report says.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

From the original article

3 min read at the-decoder.com

KPMG deployed more than 100 AI agents alongside the regular audit of a UK-listed company. The agents asked "better questions" than the audit team in some cases, according to UK head of audit Cath Burnet.

The Decoder

By Maximilian Schreiner · the-decoder.com

Read the full article at The Decoder

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.