Research links model familiarity to branded AI searches

When an assistant uses web search, it may turn one question into several smaller searches. A geoSurge study found that brands Gemini already recalled were more likely to be named in those queries. The result helps explain one part of source discovery; it does not make unfamiliar brands ineligible.

How one question becomes several searches

Say you ask an assistant which travel insurance covers skiing. With web search enabled, it may look up several parts of the question.

It breaks your question into several smaller ones and searches each of them. “Ski cover travel insurance”, “does travel insurance cover off-piste”, “winter sports policy exclusions”, and so on. It reads the results, then writes one answer out of the lot. The industry name for this is query fan-out — one question fanning out into many searches.

The interface may show only part of that search activity. Your referral analytics do not provide the assistant’s full query log. Researchers get at them by capturing what the model requests while it works.

The query choices affect which pages the assistant has an opportunity to consult.

What the study found

geoSurge took 66 buying questions from nine industries — travel, cars, finance, business software, education, restaurants, luxury, fitness, fashion — and asked each one sixty times over twelve days, between May 29 and June 9, 2026. That produced just under 4,000 answers and 13,281 fan-out searches. The searches were measured on Gemini 3.5 Flash.

Separately, they measured which brands each model already “remembered” — knew about from its training, before any web search. Then they checked whether memory predicted what got searched for.

• Brands the model remembered were searched for in 274 of 492 cases (55.7%). Brands it didn’t remember: 161 of 924 (17.4%) — roughly a threefold gap.
• Memory came in degrees. A model’s top five most familiar brands in a category showed up in searches 67% of the time. The rest of its top ten: 39%. Everyone else: 17%.
• About 31% of all fan-out searches named a specific company rather than searching the category generally. Of those, 63% named one of the model’s top-five remembered brands.

The proportions need to be read together. Named-brand searches are dominated by a handful of famous names. But they’re only about a third of the searching. The other two thirds are generic — “best travel insurance for skiing”, not “does Acme cover skiing” — leaving a substantial part of the sample focused on topics rather than named brands.

How much of this to believe

geoSurge sells AI-visibility software, so this is vendor research. The methodology is published and the sample is reasonable, but it isn’t independent.

And it’s one model, on one twelve-day window. ChatGPT, Claude and Perplexity all fan out differently, and Seer Interactive’s parallel work on ChatGPT’s fan-out patterns points the same direction without being the same measurement. Treat the percentages as a snapshot of the tested model and prompts, rather than a forecast for every assistant.

The searches nobody optimised for

Seer’s Gemini research adds the detail that changes what you’d actually do about this. Reviewing about 500 prompts and 11,029 captured fan-outs, they found 95% of the fan-out phrases have no search volume at all, and 26% contained a brand name.

No recorded search volume does not mean no relevant page exists or nobody has covered the topic. It means these exact phrases were not measured as conventional search demand in the dataset.

This supports covering the practical sub-questions behind a topic, rather than choosing every section from an exact-match keyword list. What decides those is whether a machine, running an odd phrase you’ve never heard of, finds a page of yours that answers it plainly — and can actually read that page. Some retrieval systems only see the HTML served before JavaScript runs. Check that response for the essential content instead of assuming every assistant sees the rendered browser page.

What to do about it

• Stop optimising only for questions you can see. Your keyword list is the visible layer. Also consider the related questions a reader needs answered.
• Write for the sub-question, not the headline. If your topic is ski insurance, the model is separately searching exclusions, off-piste rules, equipment cover and medical limits. Answer each of those clearly in its own section rather than burying them in one long page.
• Check the page is readable as sent. Server-rendered HTML, no crawl blocks, no critical content injected by a script after load. This is unglamorous and it is the gate everything else sits behind.
• Measure what you’re actually in, not what you rank for. Being remembered by a model isn’t something you can look up in a rank tracker. You find out by asking the engines real questions and recording who gets cited.
• Set expectations with clients honestly. A small brand may be less familiar to the tested model. Generic queries may offer additional opportunities, but the study cannot predict a particular client’s results.