Professional services firm KPMG has withdrawn a research report titled “Redefining excellence in the age of agentic AI” after multiple organizations cited in the document reported that the claims regarding their AI adoption were inaccurate. The incident, confirmed by the Financial Times, highlights the risks of using generative AI to produce corporate thought leadership without rigorous human verification.
Why were the report’s claims flagged?
The report, published in October 2025, contained false information regarding the AI integration strategies of several high-profile entities. According to the Financial Times, organizations including UBS, the UK’s National Health Service, Swiss Federal Railways, and Transport for London disputed the report’s findings, stating that the claims about their specific AI usage were either entirely untrue or significantly misleading. The research group GPTZero identified these inaccuracies, characterizing them as AI hallucinations—instances where a large language model generates plausible-sounding but factually incorrect content.
How is KPMG responding to the errors?
KPMG has removed the document from its public-facing websites while it conducts an internal review. A spokesperson for the firm stated that KPMG expects its personnel to adhere to strict guidelines regarding the responsible use of AI. This mandate includes an explicit requirement for human oversight to validate all content and verify independent sources before publication. The firm is currently investigating how the errors bypassed these editorial safeguards.
What is the broader context for AI-generated corporate reports?
This incident follows a similar withdrawal by EY in September 2025, which retracted a report on loyalty rewards programs. In that instance, the document appeared to feature fabricated footnotes and hallucinations generated by AI tools. These back-to-back cases underscore a growing tension within the professional services sector: firms are rushing to demonstrate expertise in agentic AI while simultaneously struggling to implement the quality control measures necessary to prevent AI-driven misinformation in their own publications.
Key takeaways on AI-driven corporate content
- Verification protocols: Major firms are currently reviewing internal AI usage policies to ensure human-in-the-loop verification remains mandatory for public reports.
- Hallucination risks: The incidents at both KPMG and EY demonstrate that even sophisticated generative models can produce convincing, yet entirely false, corporate data.
- Industry impact: Organizations are increasingly forced to publicly correct the record when AI-authored reports mischaracterize their technical operations or strategic initiatives.
As firms continue to adopt agentic AI—autonomous systems capable of executing tasks without constant human intervention—the reliance on these tools for drafting research poses a significant reputational risk. The current trend suggests that until AI output reaches a higher threshold of factual reliability, the demand for human editorial oversight in professional publishing will likely intensify.