AI Search

88% Disconnect Between Google Top 10 and ChatGPT Recommendations, Study Finds

A new study claims an 88% disconnect between Google top-10 rankings and conversational AI recommendations, suggesting that ranking well on Google and being recommended by an AI assistant are now largely separate achievements. The finding comes from a vendor press release, not peer-reviewed research, so it should be read as a commercial claim backed by a stated methodology, not as settled science.

The headline numbers

The headline finding is an 88% disconnect between Google top-10 rankings and conversational AI recommendations. Put the other way, the overlap between URLs cited in AI recommendations and Google's top 10 is 12.0%, according to Ahrefs Data Lab figures cited in the release. That is a small overlap by any measure, and it is the number most likely to be quoted from this study.

A second figure adds context. The study says 46.5% of AI citations come from pages ranked outside Google's top 50 organic results. In other words, nearly half of the pages AI systems cite are not just outside the top 10 — they are outside the top 50, in positions that would earn little or no organic traffic from Google in the usual sense.

A third figure puts the traffic comparison in perspective. Google sends 190 times more search traffic volume to external websites than ChatGPT, according to Ahrefs and StatCounter figures cited in the release. That matters because it reminds readers that AI recommendations, however interesting, still account for a far smaller flow of clicks to external sites than Google search does.

How the study was done

The study was carried out by Dunstan Research and distributed as a press release via ABNewswire on November 9, 2026. Its methodology, as described in the release, involved evaluating growth agencies and discovery platforms on a 100-point framework, using empirical data collected from March to October 2026.

That description leaves important questions unanswered, at least in the material supplied for this article. It does not specify which prompts were used, how many prompts were tested, which AI systems were included beyond the general category of conversational AI, or how the 100-point framework was weighted. Without those details, it is difficult to assess how far the findings generalise.

This is not a criticism of the study's authors so much as a statement about the limits of a press release as evidence. A press release is a summary designed to be read quickly, and it may omit methodological detail that would appear in a fuller report. Readers who want to rely on the findings should look for that fuller report and check whether it answers the questions above.

Why the overlap might be so low

Analysis, rather than reported fact, can help explain why an 88% disconnect is plausible even if the exact figure is uncertain. Google rankings and AI recommendations are built for different jobs: Google ranks pages for a query, while a conversational AI selects sources to support an answer, often drawing on pages that are useful as evidence rather than pages optimised to rank.

AI systems also draw on sources that Google may not rank highly for a given query, including forums, niche guides, documentation and pages that answer a specific sub-question well. A page ranked outside the top 50 for a broad query might still be the clearest source for one part of an AI answer, which would help explain why nearly half of citations come from beyond the top 50.

None of this proves the 12.0% overlap figure is correct. It only suggests the direction of the finding is not surprising. Other studies in the same area have also found low overlap between AI citations and Google top-10 results, which gives the general pattern some support even while the precise numbers remain tied to this study's own method.

Why it matters for SEOs and link builders

For SEOs, the study is a prompt to measure AI visibility separately from organic rankings, rather than assuming one stands in for the other. If only 12% of AI-cited URLs sit in Google's top 10, then a rankings report alone will miss most of what AI systems are citing. Tracking which pages are cited, and for which prompts, is a separate job that needs its own tools and its own reporting.

For link builders, the implication is that links and mentions may matter in different ways for AI visibility than for classic rankings. A link from a page that AI systems cite could carry value even if that page does not rank in the top 10, while a link from a high-ranking page that AI systems never cite may matter less for AI-driven discovery. That does not diminish links, but it does change how their value should be assessed.

The cautious takeaway is to treat the 88% figure as a vendor claim with a stated method, not as a universal law. Use it as a reason to check your own AI visibility, compare it with your own rankings, and see whether the disconnect holds for your own pages and prompts. Your own data will tell you more than any single study can.

Sources

Sources: Dunstan Research via ABNewswire press release, November 9, 2026, reporting an 88% disconnect between Google top-10 rankings and conversational AI recommendations, with data from Ahrefs Data Lab.

Related reading on this site: AI Share of Voice Is Not Organic Rankings — and the Overlap Is Small, Only 4.1% of ChatGPT's Cited Sources Rank in Google's Top 10, LoudFace Finds, Why Backlinks Alone Won't Save Big Brands in AI Search, Reddit Mentions Track AI Visibility at 0.75 — Stronger Than Any Other Signal, Ahrefs Finds, AI Visibility vs. Rankings: Are Agencies Measuring the Wrong Thing for 2027?, and High-Authority Links Lost Their AI Visibility Edge in Q3.

More from Linkbaits

Turn this playbook into links

Linkbaits helps you create content that earns citations and backlinks. Start with the toolkit and build your first linkable asset this week.

Start free trial