Listed is not picked: AI mentioned the typical brand in 19.8% of answers and recommended it in 0.2%
A report from the AI visibility vendor 3RD looked at 60 brands and 493,856 checked answers. Being on the list and being the choice turned out to be two very different jobs.
On Oct 19, 2026, 3RD published a report called "The state of AI visibility, Q4 2026". 3RD sells AI visibility tracking, and the data comes from its own customers. Keep that in mind as you read the numbers below. They are a benchmark, not a census.
One finding stood out. Across 60 brands and 493,856 checked answers, the median brand was mentioned in 19.8% of answers. It was recommended over the alternatives in 0.2%.
Who ran the study and on whom
The window was Sep 19 to Oct 18, 2026. 3RD tracked more than 18,000 prompts and 732,000 AI answers from up to 11 assistants, including ChatGPT, Gemini, Perplexity, Claude, Copilot, Google AI Overviews and AI Mode. The core sample is 111 brands from 51 customer organizations across 80+ industries.
The caveats come from 3RD itself. The sample leans Nordic, mostly Danish brands. Agencies run some of the brands. The data is observational, and nothing in it is a controlled experiment. The per-brand deep checks, including this one, use 60 brands, capped at three per customer organization.
What counts as mentioned and what counts as recommended
Mentioned means the brand appears in the answer. Recommended is stricter. 3RD counts an answer as recommending a brand only when it recommends that brand over the alternatives. Being mentioned, or listed first, does not count.
3RD checks every non-branded answer, including questions that do not ask for a recommendation. It says the definition is strict on purpose, and that this is why recommendation rates are low in absolute terms. So do not read 0.2% as "AI almost never likes brands". Read it as a gap between two measures.
The gap in the numbers
The median brand was mentioned in 19.8% of answers and recommended in 0.2%. 3RD puts that at about 1 recommendation for every 65 mentions. Of the 60 brands, 44 were recommended in under 1% of answers, and 7 reached 5% or more.
And 15 of the 60 brands were never recommended in the whole month. By prompt type, the median recommended rate was 0.3% on must-win prompts, 0.15% on core prompts and 0% on foundational prompts.
Your own site is a small part of the evidence
The same report looked at 5,926,953 citations for 100 brands. A median of 6.8% of the sources AI cited in a brand’s category pointed to the brand’s own domains. For 41 of 100 brands, AI cited competitors’ websites more often than the brand’s own.
3RD classifies sources by domain pattern, not human review, and says to read this as directional. Still, the picture is clear. Most of what AI reads about your category sits on other people’s pages.
What to do this week
These steps follow 3RD’s own advice and the numbers above. Tick them off as you go.
How Shopthru.ai helps
Shopthru.ai is an AI agents team for consumer brands. It asks ChatGPT, Gemini, Google AI, Perplexity and Copilot what your shoppers ask, finds where AI gets your products, prices or policies wrong, and drafts each fix into an approval inbox. It ships only on your yes, then checks again. Instagram and Amazon are next.