KPMG pulls agentic AI report after companies challenge fabricated claims

London: The global consulting giant KPMG has withdrawn its report, “Redefining Excellence in the Age of Agentic AI”, after several organisations named in the study challenged claims about their supposed use of advanced AI systems. According to a report by the Financial Times, the publication contained multiple fabricated case studies and inaccurate descriptions of how companies were deploying so-called “agentic AI” technologies.

The controversy centres on a series of success stories presented as real-world examples of organisations using autonomous AI agents to perform complex tasks. However, many of the companies featured in the report said the examples were either exaggerated or entirely false, raising concerns that the content may have been generated through AI hallucinations and insufficiently checked by human reviewers.

The inaccuracies were first flagged by researchers at GPTZero and later independently verified by the Financial Times. Once alerted, several prominent organisations demanded corrections or removal of the report, prompting KPMG to take the publication offline while it investigates how the errors occurred.

Among the most notable objections came from UBS. The report claimed the bank had integrated AI agents across its investment advisory and risk management operations through a custom platform built with Microsoft technology. The bank swiftly rejected the assertion, with a spokesperson describing the claims as “factually incorrect” and requesting their removal.

Similarly, Swiss Federal Railways was cited as using AI agents to help travellers plan and book journeys based on live travel conditions and environmental considerations. The railway operator later stated that the description was inaccurate and did not reflect its actual operations.

The report also claimed that Transport for London was using AI agents to predict congestion patterns and coordinate transport flows across the city. Transport officials reportedly described the characterisation as misleading.

Meanwhile, NHS Greater Manchester was presented as using AI agents to organise patient information, automate referrals and predict hospital readmissions. Representatives from the healthcare body said those claims did not align with reality.

The incident has reignited concerns about the growing use of generative AI in corporate research and publishing. It also follows another high-profile embarrassment in the consulting industry. Last month, rival firm EY reportedly withdrew a major study after GPTZero identified fabricated references and other AI-generated errors.

Speaking about the broader implications, GPTZero chief executive officer Edward Tian warned that such mistakes can have consequences beyond the original publication. He noted that before KPMG removed the report, several technology publications and even a major European newspaper had already cited its findings, allowing questionable information to spread further.

KPMG has since launched an internal investigation into the episode. A company spokesperson said the firm takes the “accuracy and integrity” of its published content seriously and acknowledged that internal guidelines on responsible AI use may not have been followed.

The spokesperson added that KPMG expects employees to apply human oversight when using AI tools, including validating content and independently verifying sources before publication.

The episode serves as a timely reminder that while AI can help write reports at speed, credibility still depends on human judgement. In a business built on trust, getting the facts right remains a task no machine can fully automate.