Ask three AI engines the same ten shopping questions and you get three different reading lists, but one domain shows up everywhere. A Beket AI study run on November 15, 2026 found RTINGS.com, the lab-testing review site, was cited 21 times across ChatGPT, Perplexity and Google AI Overviews, more than any other domain. The study logged 252 citation events in total, and only three domains were cited by all three engines.
Beket AI asked ChatGPT, Perplexity and Google AI Overviews ten unbranded buying questions as a US, English-language user, then logged every source each answer cited. Of the 30 possible answers, 28 were recorded, because AI Overviews answered only 8 of the 10 questions. The 252 citation events split unevenly: Perplexity cited 117 sources, ChatGPT 72, and AI Overviews 63, across 124 distinct domains. RTINGS took 21 of those citations, or 8.3% of the total, and appeared for five of the ten questions: robot vacuums, running shoes, college laptops, VPNs and mattresses.
Behind RTINGS came YouTube with 13 citations and PCMag with 11. The overlap between engines was strikingly small. Only three domains, RTINGS.com, TechRadar and Wired, were cited by all three engines. That means the engines are not drawing on one shared pool of trusted sources for shopping questions; each has its own reading list, and a site that is visible in one engine can be invisible in another.
The split by source type is where the study gets useful. Perplexity sent readers mostly to publishers: 92 of its 117 citations, or 78.6%, went to publisher sites, with only 24 going to brand sites. ChatGPT was nearly the opposite, sending 40 of its 72 citations, or 55.6%, to the brands themselves, against 30 to publishers. Google AI Overviews was the most scattered: only 25.4% of its citations went to publishers, 23.8% to brand sites, 28.6% to user content and 22.2% to other sources, with YouTube alone taking 13 of its 63 citations.
That pattern matches what the engines are built to do. Perplexity behaves like a research tool that leans on independent reviews. ChatGPT behaves more like a shopping assistant that goes straight to the seller. AI Overviews behaves like Google search itself, pulling in video, forums and its own properties alongside publishers. For a brand, the implication is that the same buying question can surface your own product pages in ChatGPT, a review of your product in Perplexity, and a YouTube video in AI Overviews, all at once.
The study is a reminder that there is no single AI search visibility to optimize for. A link or mention that earns a citation in Perplexity may do nothing in ChatGPT, and a product page that ChatGPT cites directly may never appear in AI Overviews. The winning play is to cover all three surfaces: lab-tested, data-rich review content of the kind RTINGS produces for Perplexity, clear and complete product pages for ChatGPT, and video and community content for AI Overviews.
It also shows what kind of third-party coverage actually gets cited. RTINGS won not with opinion but with measured, repeatable testing that the engines can lean on as evidence. For link builders, that is the model: a placement on a site that tests, measures or documents something is worth more in AI search than a placement that only praises. Beket AI itself notes the limits, that ten questions on one day are a narrow window, and that it also checked the pattern against 166 stored audits. Even so, the direction is clear, and it rewards the sites that do original work.
Watch whether the pattern holds beyond shopping questions. Beket AI's own city studies from the same week found a different shape for local services, where directories and official bodies dominated, with the New York City Bar Association taking 23 citations for lawyer questions and Clutch taking 69 for Los Angeles marketing agencies. The most-cited source depends heavily on the question type, so track the engines separately for each market you care about, and re-check often, because these answers change by the day.
Sources: Beket AI, November 15, 2026.
Related reading on this site: Wikipedia Leads ChatGPT Citations at 10.1% in Ahrefs\' November Ranking, Reddit Takes 29.9% of Gemini Citations in Ahrefs\' November Ranking, Amazon Takes 16.1% of Copilot Citations in Ahrefs\' November Ranking, AI Overview Citations: What 15 Studies Agree On, and Don't, Only 2.9 Percent of AI Citations Point at the Brand\'s Own Domain, and Perplexity's Citation Funnel: 29 Query Variants, 10 Pages Retrieved, 3 Cited.
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