Four Engines, Four Different Games
For most of the last two years, the working assumption in AI visibility has been simple: rank well on Google, get cited by AI. A new benchmark shows that assumption only holds for half the platforms that matter.
CiteLens ran 320 real buyer queries across three consumer sectors through ChatGPT, Perplexity, Claude, and Google AI Mode, then checked every citation each engine produced against that same query’s Google and Bing top-10 results. The headline finding, in the founder’s own words: “SEO gets you into Google’s AI and Perplexity, but it barely moves ChatGPT. Marketers are told to optimize for AI as if it’s one thing — our data shows it’s at least three.”
Who Cites What
The overlap between each platform’s citations and Google’s own top-10 results:
- Google AI Mode — 93% overlap (correlation coefficient 0.92)
- Perplexity — 89% overlap (correlation coefficient 0.87)
- Claude — 53% overlap
- ChatGPT — 30% overlap
Google AI Mode and Perplexity behave almost like a re-skinned version of the search results page — unsurprising for AI Mode, since it lives inside Google Search, and consistent with Perplexity’s own heavy reliance on live web retrieval. Claude sits in the middle. ChatGPT is the outlier, and not by a small margin.
Ready to Monitor Your AI Visibility?
Track how AI chatbots mention your brand across ChatGPT, Perplexity, and other platforms.
Why ChatGPT Diverges
The obvious explanation — that ChatGPT is just pulling from a different search index — doesn’t hold up. The same study found ChatGPT’s citations matched Bing’s top 10 results less than 4% of the time, ruling out “it’s just using Bing instead of Google” as the answer.
What correlates instead is entity authority : how consistently and credibly a brand is described across the web, independent of whether any single page ranks. One clean signal of that split shows up in Wikipedia-backed citations.
Claude traced 58% of its citations back to a Wikipedia-referenced source. ChatGPT traced only 21%. Both numbers point the same direction as the ranking-overlap data: Claude and ChatGPT are weighting broader entity credibility over search-engine rank, they just weight the specific sources differently.
Zero-Ranking Pages
A separate, larger-scale analysis from Ahrefs backs this up from another angle entirely. Looking at ChatGPT’s 1,000 most-cited pages, Ahrefs found that 28.3% have zero organic visibility on Google — no ranking keywords at all, in a dataset otherwise averaging 279 ranking keywords per page.
Some of that gap is timing — content too new to be indexed yet. Some is topic — long-tail subjects with real relevance but little search demand. But the pattern is the same one CiteLens found: for ChatGPT specifically, a meaningful share of what actually gets cited was never going to show up in an SEO rank tracker in the first place, because ranking was never the mechanism that put it there.
Stay Updated on AI Visibility Trends
Get the latest insights on AI mentions, brand monitoring, and optimization strategies.
What Predicts Citations
Put the two studies together and a clear split emerges, matching what we’ve also found looking at why AI Overviews lean so heavily on top-10 results :
Ranking-driven platforms — Google AI Mode, Perplexity. Classic technical and content SEO still does most of the work here. If a page ranks, it has a strong chance of being cited. Investment in how LLMs actually select citations pays off most directly on these two.
Entity-driven platforms — ChatGPT, Claude. Ranking position is a weak predictor. What correlates instead is consistent brand presence across the wider web: Wikipedia, review platforms, forums, and credible third-party mentions that build recognition independent of any one page’s search performance. Tracking Claude specifically requires this same lens, since a strong Claude AI optimization approach looks meaningfully different from a Perplexity-first playbook.
What To Do Now
Run one strategy for both groups and you’ll under-invest in one of them:
- Keep the SEO fundamentals sharp for Google AI Mode and Perplexity — they’re still the closest thing to classic search optimization inside AI.
- Build entity credibility deliberately for ChatGPT and Claude — earned Wikipedia coverage, genuine reviews, and consistent facts about your brand across independent sources.
- Stop reading one citation rate as “your AI visibility.” A platform-by-platform breakdown is the only way to see which lever is actually working, and which one you’re neglecting.
- Re-run the check periodically. These correlations reflect where each platform’s retrieval approach stands today; as models and their grounding methods change, the split can move.
The Bottom Line
“AI SEO” isn’t one discipline, and treating it like one means over-investing in the platforms where ranking already works and under-investing in the ones where it doesn’t. Knowing which is which for your own brand, on every platform that matters, is exactly what Am I Cited is built to show you.
