Does being talked about extra on the net result in extra mentions in AI assistants? I’m defining mentions because the variety of internet pages that point out an entity no less than as soon as.
I regarded on the high 50 web sites talked about in Ahrefs Model Radar for Google AI Overviews, ChatGPT, and Perplexity.
That is throughout ~76.7M AI Overviews, 957k ChatGPT prompts, and 953.5k Perplexity prompts for the month of June 2025.
Let’s dig in.
I’ll give the standard correlation doesn’t equal causation disclaimer.
Google has a robust correlation, the place it looks like branded internet mentions result in extra visibility. The others didn’t correlate as properly.
We noticed practically the identical correlation for AI Overviews throughout 75,000 web sites. In actual fact, Branded internet mentions had been the best correlating issue. Right here’s the info from that research.


Nevertheless, Perplexity and ChatGPT didn’t correlate practically as a lot. Perplexity had a weak correlation and ChatGPT a really weak correlation.
Google has favored manufacturers for a very long time, and the opposite AI assistants could not have the identical built-in techniques for it that Google does. Google favors manufacturers to battle false info, to lean into trusted content material. Google’s former CEO Eric Schmidt was as soon as requested concerning the web changing into a cesspool the place false info thrives. He mentioned:
Manufacturers are the answer, not the issue.
Listed here are the Spearman rank correlations for branded internet mentions vs. AI visibility throughout the highest 50 domains in every system:
| AI Assistant | ρ (Mentions vs Visibility) | Correlation |
|---|---|---|
| Google AI Overviews | 0.65 | Sturdy |
| ChatGPT | 0.15 | Very Weak |
| Perplexity | 0.30 | Weak |
Remaining ideas
We’ll run this on a bigger pattern of websites and see if something modifications, however contemplating how properly the outcomes for Google AI Overviews lined up with the bigger research, I’m undecided they are going to change a lot.

