Because the sources AI reads to answer a buyer question, roundups, review sites, Wikipedia, mostly already name the big brands. In Reflexa's own Index, engines named 494 project-management tools, yet three took 38% of the first choices (Reflexa AI Visibility Index, October 2026).
This is the same pattern AI search data keeps showing across categories: AI answers concentrate around a small set of names, and the gap between a crowded market and a short AI shortlist is wider than most owners assume.
Because an engine builds its answer from whichever sources it can read, trust and summarize fast, and those sources already agree on a handful of names. A roundup article, a review site's category page and a Wikipedia entry each tend to list the same three or four market leaders, because that is who has the most reviews, the most backlinks and the longest Wikipedia page. The engine is not independently judging every tool in a category; it is reflecting what its sources already say, and its sources already agree with each other.
Why it matters: a company absent from those same roundups and review pages starts the question already behind, no matter how good the product is.
Very: in Reflexa's own project-management Index, engines named 494 different tools, but the field narrows fast after that. Of those 494 names, only 5 were shortlisted in at least 1 answer out of 10 across the Index's 900 buyer-question answers, and just 3 of those, Asana, monday and ClickUp, took 38% of the first choices between them.
| Layer | What it measures | Observed (project management, Oct 2026) |
|---|---|---|
| Named at all | Tools that appear anywhere in an answer | 494 tools |
| Real shortlist | Tools shortlisted in 10%+ of 900 answers | 5 tools: Asana, monday, ClickUp, Trello, Wrike |
| First choice | Share of first choices taken by the top 3 | 38% (Asana, monday, ClickUp) |
Proof: Reflexa AI Visibility Index, project-management edition, October 2026.
Bottom line: most of the 494 named tools are mentioned once in passing; the real competition for a buyer question happens among a handful of names.
No, named, shortlisted and first choice are three different numbers that move independently, and the gap between them is where most of the confusion sits. In the same Index, Asana is named in 49% of the 900 answers, shortlisted in 41%, and the engine's literal first choice in only 19%. Trello shows the opposite kind of gap: it is named in 29% of answers, but its own site is cited as a source in only 4% of them, evidence that AI learned about Trello from other pages, not from Trello's own site.
Do this: track all three numbers for your own brand, named, shortlisted and first choice, instead of collapsing them into one score. A brand can be named often and still lose the actual pick.
On the specific question, not the generic one. "Best project management tool" almost always returns the same 3-5 names, because that question matches what every roundup already ranks for. "Best project management tool for a 5-person agency on a $20/month budget," or "for a construction crew that needs it on a job site," is a narrower question, and the generic market leader is often not written for that case. A page that answers the specific version in its first sentence can be the clearest, sometimes the only, source an engine finds for it.
Why it matters: this is also the reason a small or local business can still get recommended; see Is AI visibility worth it for a small business? for what is realistic to fix in a week versus what takes longer.
None of this requires a budget to start:
Bottom line: you will not out-rank the category leader on the generic question this quarter; you can realistically own the specific one.
Because ChatGPT, Claude and Perplexity build an answer from roundups, review sites and Wikipedia, and those sources already repeat the same handful of names most often. The engine is reflecting what its sources say, not judging your product.
In Reflexa's own project-management Index (October 2026), engines named 494 different tools across 900 buyer-question answers, but only 5 made the shortlist in at least one answer out of ten, and 3 of those took 38% of the first choices.
No. In the same Index, Asana is named in 49% of answers, shortlisted in 41%, but the engine's literal first choice in only 19%. Named, shortlisted and first choice move independently, so track all three, not one score.
On specific questions, not generic ones: for a particular budget, team size, city or use case, where the market leader is not actually the best answer and a smaller company's own page can be the clearest source.
Stop competing on the generic question the market leader already owns, and build pages that answer the specific question your actual buyer asks. Then confirm AI can read, trust and cite that page with a free check.
The free check reads your pages the way ChatGPT, Claude and Perplexity read them and shows what stops them, with the evidence behind every finding. 3 minutes, evidence included.
See the same numbers for your site: every AI visit counted per page, beside Search Console · free first check · på svenska