One of the world’s most prestigious consulting firms published a report about the power of AI. The irony is that AI may have been the thing that made the report worthless.
KPMG’s Total Experience: Redefining Excellence in the Age of Agentic AI, published in October 2025 as part of its Global Customer Experience Excellence study, has come under fire after an investigation revealed that the vast majority of its citations were fabricated, misattributed, or unverifiable. The report included case studies about major firms like UBS that showed signs of exaggeration or outright fabrication.
The numbers are brutal
AI-detection firm GPTZero conducted an analysis of the report in June 2026, scrutinizing all 45 citations. The results were not subtle.
Only 5 citations were fully accurate. Five out of forty-five. That’s an 11% accuracy rate.
Of the remaining citations, 28 contained paraphrased or fabricated components, including made-up titles and fictional authors. The other 12 were too vague to even verify.
Around half of the claims in the entire report were deemed either fake or misattributed. The Financial Times highlighted bogus case studies that exaggerated AI implementation benefits at companies including UBS and various health providers.
A pattern across the Big Four
KPMG isn’t alone in this particular embarrassment. The consulting industry appears to have a growing AI-generated hallucination problem.
EY withdrew a report in May 2026 after discovering AI-generated citation errors. Deloitte Australia has faced similar concerns regarding nonexistent references appearing in its materials.
Courts have sanctioned lawyers for filing briefs with AI-fabricated case citations. Academic journals have flagged papers with nonexistent references. But the consulting sector presents a unique risk because its entire business model rests on the credibility of its analysis.
What this means for anyone relying on consulting research
Companies routinely use Big Four research to justify strategic pivots, technology investments, and board-level decisions around AI adoption. If the case studies showing transformative AI results at firms like UBS were exaggerated or fabricated, then the executives who used those case studies to build their own AI business cases were working from a fictional foundation.
If three of the Big Four have now been caught with AI-generated errors in published materials, the reasonable assumption for any client is that unverified AI-assisted content may be lurking in other reports they’ve already consumed and acted on.
KPMG published a report about AI excellence that was undermined by AI’s most basic and well-known failure. That’s not just a quality control lapse. It’s a product demo gone wrong in the most public way possible.