Why ChatGPT recommends some brands and not others: the lists AI reads
Yorker Media Editorial · 28 October 2026 · 5 min read
A growing share of buyers now start with a question to ChatGPT, Perplexity, Gemini or Google's AI answers: "What's the best CRM for a 10-person agency?" The answer names three to five products. If yours is not one of them, you are not on the shortlist, and you will rarely find out why.
This guide pulls together what the 2026 studies say about where those answers come from, and turns it into a plan you can run this month.
What AI assistants actually read
Every assistant that browses the web leans on the same kind of page when it recommends software: the ranked list. In an analysis of AI Mode, ChatGPT and Perplexity citations, listicles were the most-cited page type at 21.9%, ahead of articles (16.7%) and product pages (13.7%); for B2B SaaS questions, ChatGPT cited listicles 18.8% of the time (summary of six 2026 studies).
The second pattern is just as strong: assistants trust other people's pages about you more than your own. Brands are about 6.5 times more likely to be cited through third-party sources than through their own sites, and in Ahrefs' study of 75,000 brands, branded web mentions correlated with ChatGPT visibility at 0.664, against 0.218 for backlinks. Correlation is not proof of cause, but the direction is consistent across studies (same summary as above).
In plain terms: the pages that decide your AI visibility are mostly not yours.
Why your own "best X" page is not enough
Many vendors have noticed this and published their own "10 best tools" pages, with themselves at #1. A study of 18,084 software questions found that for "alternatives" questions, 34–47% of the listicles AI tools cited were written by a vendor about its own category; 87% of those put the vendor first, and only 16% disclosed it (Minuttia, October 2026).
It works less well than it looks. When Google's AI Overviews cited a brand's own listicle, it left that brand out of its actual recommendation 69% of the time, in a three-month study of 100 B2B "best software" queries (reported in the same 2026 roundup). Being read is not the same as being recommended. An independent page that ranks you is a stronger signal than a page where you rank yourself.
Breadth beats a single placement
The assistants do not read the same sources. Across 680 million citations, only 11% of domains were cited by both ChatGPT and Perplexity (AuthorityTech audit). A single article on one site gives you one chance with one engine. Places on many independent lists, across many sites and niches, give every engine something to find.
Most teams know this and still have no plan: in a 2026 survey of 169 B2B SaaS marketers, 93% said AI visibility matters, and only 14% had a strategy for it (CommonMind).
A five-step plan for this month
1. Write down the questions your buyers ask. Ten to twenty prompts, in their words: "best X for Y", "X alternatives", "X vs Y". Run each in ChatGPT, Perplexity, Gemini and Google, and note which products come up and which pages are cited.
2. Find the lists behind the answers. The cited pages are usually ranked lists on review and comparison sites. Collect them, plus the lists that rank on Google for the same questions. These are the pages that shape the answer.
3. Get into those lists, as high as you can. A place in the top five of an independent, maintained list is the single most direct move. You can earn it through reviews and outreach, or buy it where it is sold openly. On the Yorker Media marketplace you can search 63,000+ ranked lists across 35 independent sites and take #1 to #5 in any of them, monthly, yearly or once for good.
4. Keep your facts identical everywhere. Assistants cross-check. If your pricing, plan names or key features differ between your site and the lists, you give them a reason to leave you out. Send every list owner the same short fact sheet, and update it when anything changes.
5. Check again every month. Re-run your prompts. Note where you appear, which pages are cited, and where a competitor replaced you. Treat it like rank tracking: the answers move.
What no one can promise
No publisher, agency or tool can guarantee that an assistant will name you; the models change their sources and their wording often. What you can control is how many credible, independent pages describe you accurately, and how high you sit on them. That is what moves the odds.
Frequently asked questions
Is GEO different from SEO? Mostly in where the effort goes. SEO rewards your own pages ranking; GEO (generative engine optimization) rewards being described well on other people's pages that assistants read, especially ranked lists, reviews and comparisons.
Do backlinks still matter? They still help search rankings, and the pages that rank are the ones assistants find. But in the studies above, brand mentions on third-party pages track AI visibility more closely than backlinks do.
How many lists should I be on? Start with the lists cited for your top five buyer questions, then widen to every independent list in your category. Because each assistant reads different sites, spreading across several sites beats a single placement.
Keep reading
What a #1 spot on a 'best software' list costs in 2026, and how to tell if it pays
Real prices for #1 to #5 on independent best-of lists, monthly, yearly and once, and a simple way to work out whether a placement will pay for itself before you buy it.
Link insertion, guest post or a ranked spot: which one to buy, and when
Three ways to appear on an independent site, three different jobs. What each one gets you, what it costs, and how to choose between them for SEO, AI visibility and sales.