If you want your own figure, the segment worth building is narrower than most people set up. Compare assistant referrals against branded organic search rather than against all organic, over at least a quarter, and exclude any campaign traffic. It will be a small sample and it will be about your audience, which makes it more useful for your decisions than a published study about somebody else's.
Absence is not disqualifying on its own, since their category is crowded and they may serve a niche. But they should have an interesting answer, and the answer should not be defensive. A practitioner who has run this test on themselves will have thought about it and will tell you what they found.
What We Genuinely Do Not Know Several things are worth admitting rather than papering over. We do not know how the systems weight their signals against each other. We do not know how much residual influence training data has once retrieval is involved. We cannot reliably distinguish a change in your visibility from a change in the model's behaviour.
The defensible version states the mechanism, cites the available evidence with its sample sizes, presents your own segmented data however thin, and is explicit that most of the channel's value is not measurable through referrals at all.
What to Build and What to Buy Build the prompt set and the measurement habit internally. They are cheap, they depend on knowledge of your customers that no agency has, and owning them means you can audit anyone you hire.
So the work splits in two. Make your own pages quotable, which you control entirely, and get recommended by ai accurately represented on the pages that already get cited, which you control only partly. Neither half works alone.
The pattern is consistent across most categories. Review platforms, industry publications, documentation, forum threads and comparison articles appear far more often than brand websites. When a brand site is cited it is usually a specification page, a pricing page or a technical document rather than a homepage or a landing page.
This is why glossary style content and plainly written explainers appear so often. It is also why leading with the answer matters so much: a page that spends four paragraphs arriving at its definition contains nothing usable until the fifth.
What It Is Doing Under the Hood Simplified, the sequence runs like this. Your question is rewritten into one or more search queries. Results come back. A subset of pages is fetched and read. The model composes an answer from what it read and attaches citations to the specific claims it lifted.
One further caution applies to how this gets used in a pitch. An agency quoting a conversion multiple without its sample size is either unaware of the provenance or hoping you are, and both are informative. Asking where a number came from is a reasonable question that costs nothing, and the quality of the answer tells you a good deal about how your own reporting will be handled.
A small habit pays off here more than it should. Give each substantial page a short section that states the plain facts in one place: what the thing is, what it costs, how long it takes, who it suits and who it does not. That block is disproportionately likely to be the passage that gets lifted, because it answers several likely questions in a form that survives extraction, and it costs almost nothing to add to a page you were writing anyway.
Nineteen properties can show a real trend and cannot support a confident statement about the market. When that number is repeated without its sample size, as it usually is, it stops being evidence and becomes a slogan.
What Ranking Does and Does Not Buy You Ranking still helps, because the retrieval step usually starts with a search. But it buys far less than people assume. Ahrefs examined 15,000 long-tail prompts across four assistants in July 2025 and found roughly 80 percent of cited pages did not rank for the original query at all, with about 12 percent in the top ten.
The reasonable reading is that ranking gets a page considered while quotability and corroboration decide whether it is used. Treating a strong search position as an entitlement to appear in answers is the mistake that catches out established brands most often.
This matters more than any subtlety about model training. It means recommendations are built largely from pages that exist right now, which is why a page published this month can influence an answer this month, and why a brand absent from the retrievable web is absent from the answer regardless of how well known it is offline.
Two caveats belong next to that number every time it is used. Opollo sells services in this space, so it is vendor research and interested. And business to business brands are not representative of retail, local services or consumer products.
A reasonable rule for planning a content programme is to publish fewer pages and maintain them properly. Twenty pages carrying current figures will out-earn a hundred that were correct on the day they shipped, because freshness is weighted and stale specifics actively cost you. Most teams discover this by building the hundred first, then finding they cannot review them and quietly letting the whole set go out of date.