KPMG pulls AI report after UBS, NHS and others say the case studies were invented

KPMG has withdrawn its report "Rethinking Excellence in the Age of Agentic AI" after four major organizations from different countries and sectors publicly disputed the descriptions of their AI projects included in the document.

What happened to the report

The KPMG report targeted a corporate audience and focused on best practices for deploying agentic AI, meaning systems that can carry out multistep tasks autonomously. It featured case studies from large companies and government bodies, presumably as examples of successful work with the new technology. Studies like this from "Big Four" consultancies are traditionally treated as an authoritative reference when companies decide whether to adopt a technology.

The problem surfaced when the Financial Times contacted the organizations named in the report. Company representatives told the paper they had learned of their participation only after publication, and were extremely surprised. The case studies matched neither the organizations' actual projects nor their official positions on artificial intelligence.

After the wave of denials, KPMG withdrew the report without detailed comment. The company has not publicly explained how inaccurate information about specific organizations ended up in an official document.

Who disputed their case studies

Four organizations from entirely different sectors lodged complaints with KPMG:

• UBS: one of the world's largest banks said the AI initiatives described have nothing to do with it
• NHS (the UK's National Health Service): said its case was presented inaccurately or distorted
• Swiss Federal Railways: did not recognize the AI projects attributed to it
• Transport for London: also denied the description of its AI activity in the report

Notably, all four organizations came forward independently, on their own initiative and without coordination. That rules out the explanation of an isolated misunderstanding. It is also telling that none of them questioned individual details: each said the described cases simply do not relate to it. That is a fundamental difference between an inaccuracy and outright invention.

Hallucinations in a report about AI

It is hard to miss the irony. A report devoted to "excellence" in the age of agentic AI itself suffered from what remains the central problem of modern language models: hallucinations, the confident reproduction of nonexistent facts, names, quotes and descriptions. If AI tools were used to collect or summarize the case studies without proper checking, that would explain the result: convincingly written but entirely fabricated success stories. Because hallucinations sound plausible and contain specific details, they are extremely hard to spot without careful fact-checking of every claim.

"The description of our work with AI does not correspond to reality," the

Financial Times quotes representatives of several of the named organizations as saying.

The scandal hurts KPMG's reputation as a source of reliable analysis. "Big Four" reports are generally considered vetted, which is why they are cited in corporate strategies and in board presentations. If such reports can contain invented case studies, that changes how far they should be trusted without independent verification.

What this means

The KPMG incident is a symptom of a broader problem. The hype around agentic AI puts enormous pressure on analytics teams to publish material faster than they can verify the data. For any organization, it is a reminder to watch closely how its name and projects are presented in industry reports and AI marketing. Distorted facts that land in a respected report once quickly take on a life of their own, as people cite and reference them. Organizations need to react before that happens.