API
< Back to all clusters
[BUSINESS] · United Kingdom · 5 sources

started · updated

KPMG tests AI agents in audit experiments

KPMG has conducted experimental audits using AI agents alongside human teams, finding that the technology could sometimes ask better questions and perform deeper investigations into financial data than human auditors. During a trial involving an anonymous FTSE 250 company, over 100 AI agents independently collected and analyzed data, drafted conclusions, and proposed recommendations.

Despite these capabilities, KPMG emphasized that human oversight remains critical. In instances where AI agents failed to correctly match regulatory requirements with company documents, human intervention was required. The firm plans to expand this initiative, known as Project Exponential, to up to 20 larger audits.

The UK regulator, the Financial Reporting Council (FRC), has issued guidance regarding the use of AI in auditing. The FRC noted that while technology can enhance data analysis, auditors must maintain independence and remain capable of explaining and questioning AI-generated results to ensure professional skepticism.

Entities

Cath Burnet · FTSE 250 · Financial Reporting Council · KPMG · PwC