Case studies
Anonymised, with the rivals named. Every figure comes out of a run you could reproduce with the tool on this site in about a minute.
A top-30 UK accountancy practice was named in two answers out of thirty
We put 6 questions its own buyers ask to 5 answer engines, in August 2026. That is 30 answers. The firm was named in 2 of them and cited as a source in 1.
It was not named once in three of the five engines
- ChatGPT: named in 0 of 6
- Claude: named in 0 of 6
- Gemini: named in 0 of 6
- Google AI Overviews: named in 1 of 6
- Perplexity: named in 1 of 6
Its own competitors were named instead
Not the Big Four, which the engines reach for regardless of who is asking. These are firms from the practice’s own competitor list:
- Grant Thornton: named in 10 of 30
- RSM UK: named in 5 of 30
- BDO: named in 4 of 30
Of the 9 competitors the firm itself would name, the engines surfaced 8.
Why it was missing
- Proof. Rivals publish named data and tracked reports the engines can quote. This firm publishes claims.
- Topical authority. Thin coverage of the current subjects buyers are actually asking the engines about.
- Positioning. Stated less clearly than rivals, so the engines cannot confidently categorise the firm.
Subject anonymised. Rivals named, because naming a firm as winning in AI answers is a factual statement about public data. Figures unedited.