KPMG pulls AI report after companies flag ‘made-up’ case studies

KPMG has pulled a global report on agentic AI after several organisations challenged claims made about them, raising fresh questions over how consulting firms are using artificial intelligence to produce research and thought leadership.

The report, titled Redefining Excellence in the Age of Agentic AI, claimed that companies and public bodies were already using AI agents to transform operations across finance, healthcare and transport. But according to the Financial Times, several case studies in the report were found to be inaccurate and appeared to have been generated by AI hallucinations.

The errors were first flagged by tech research firm GPTZero and later verified by the Financial Times. Following this, organisations including UBS, the UK’s National Health Service, Swiss Federal Railways and Transport for London asked for the claims to be removed.

KPMG had claimed that UBS was using AI agents across investment advisory and risk management through a custom Microsoft-built platform. UBS disputed the claim, with a spokesperson calling it “factually incorrect”, according to the Financial Times.

Swiss Federal Railways also pushed back against claims that it was using AI agents to help passengers plan and book journeys based on real-time travel conditions and carbon emissions. The rail operator said the description was “not accurate”.

Transport for London said claims that it used AI agents to predict congestion and coordinate the city’s transport network were “misleading”. NHS Greater Manchester also challenged KPMG’s description that it used AI agents to organise patient records, automate referrals and predict hospital readmissions, saying it “doesn’t really align” with reality.

The episode comes weeks after EY withdrew a study that reportedly contained fake footnotes and AI-generated errors, also uncovered by GPTZero.

KPMG International has removed the report and started an internal investigation. A spokesperson said the firm takes the “accuracy and integrity” of its publications seriously and that employees may have breached internal AI-use policies.

“We expect all our people to follow our guidelines on the responsible use of AI, including human oversight to validate content and verify independent sources,” the spokesperson said.