Vendor reportAhrefs, 11 Aug 2025
AI assistants choose citations differently from Google or Bing, often picking pages from related queries rather than the exact same prompt
Citation & visibility tracking
- Reader Note:
- This is the article's own explanation for its numbers, not a measurement. The piece introduces it as 'we have a couple of theories as to why that might be' and describes query fan-out and Reciprocal Rank Fusion as mechanisms it believes are at work, with a worked example it made up rather than observed. Personalisation is offered as a second theory.
- Evidence:
- Expert judgment
- Vendor interest:
- Moderate — This comes from a company with a commercial interest in this market, so it is not a neutral source.
- Sample:
- 15,000 long-tail queries run in Google and Bing, then the same questions asked of ChatGPT, Gemini, Copilot and Perplexity; every visible cited URL was extracted and matched against search ranking positions, recorded as top 10 or top 100. Data collected in early July 2025, published 11 August 2025 — the article itself warns that results may have changed. The headline averages are unweighted means across five citation surfaces, not four assistants: ChatGPT is counted twice (in-text citations and bottom-of-response references), so it carries double weight, Perplexity carries single weight, and no surface is weighted by how many citations it actually produces. 12% is (28.6 + 8 + 6.1 + 8.6 + 8.2) / 5; the 11% Google-and-Bing figure divides the two engine totals by 10; the Bing 10% is (16.6 + 14 + 8.1 + 8.1 + 3.3) / 5. The 80% non-ranking figure is the author's own approximation, not a computed statistic.
Good for
Hearing how someone who knows the field reads it, and deciding what to check yourself.
Not for
- Anything that needs a measurement. Nobody counted or tested this.
- This comes from a company with a commercial interest in this market, so it is not a neutral source.
- The source says how much it looked at, but not how any of it was chosen.




