AI search news ·
Does topical focus make your brand more visible in AI search? Cited almost anywhere — named only where you have depth.
Kevin Indig's Growth Memo study published 2026-08-03 (syndicated by Search Engine Land on 2026-08-05) puts large-scale numbers on a split we measure constantly: being cited by an AI engine and being named by it are different events with different causes. Across 283,215 citations and 76,493 named-brand mentions from Semrush's US ChatGPT data (January–June 2026, 1,094 categories × 5 prompt variants), brands get cited far outside their expertise — but get named mostly inside it: 74% cited and 44% named in close categories, versus mentions dropping to 25% in distant ones.
Primary source: Kevin Indig — Growth Memo
The study
Indig mapped 1,458 brand entities against Semrush's US ChatGPT AI Visibility Toolkit data — 1,094 categories with 5 prompt variants each, January through June 2026, yielding 283,215 citation observations and 76,493 named-brand-mention observations — and scored every brand appearance by how topically close the category sits to the brand's core expertise, controlling for authority score, organic traffic and branded search demand. The unit distinction is the one that matters: a citation is your domain in the sources list; a mention is the engine writing your name into the answer.
The numbers
| Where the brand appears | Cited | Named mention | Both |
|---|---|---|---|
| Close categories (core expertise) | 74% | 44% | 34% |
| Distant categories | 50% | 25% | — |
Citation-only presence is nearly identical either way — 40% in close categories, 41% in distant ones — meaning engines will happily pull your page into a sources list on almost any adjacent topic. What moves with topical focus is the mention. Depth compounds it: appearing in just 1 of a category's 5 prompt variants is associated with lower mention share (−0.051); covering all 5 turns the association slightly positive. The citation-side association rises from +0.012 at 1-of-5 to +0.062 at 5-of-5. And the effect is not uniform by industry: in finance the citation-breadth association climbs from +0.054 to +0.139 and in real estate from +0.021 to +0.135 — while in legal (−0.058) and healthcare (−0.037) the mention-breadth association stays negative even at full 5-prompt coverage.
Why it matters
- The mention is the conversion event, and it is the scarce one. A customer reading an AI answer acts on the names in the text, not the domains in the source list. Indig's data says names are earned through repeat topical depth; our own corpus says the same thing from the other side — Semrush's 126M-prompt dataset showed brand visibility is overwhelmingly mentions, not citations, and in our audits AI engines recognise businesses by name in 98.8% of answers yet name them unprompted in only 19.8%.
- Scattered citations are not a strategy. If your pages get cited across distant topics at 41% while your brand goes unnamed, you are supplying the answer without being the answer. The playbook the data supports is narrow-and-deep: cover the full spread of buyer questions in your core category before touching adjacent ones.
- Legal and healthcare need their own measurement. In the two industries where depth fails to buy mentions, assuming the generic GEO playbook works is exactly backwards — measure whether the engines name your firm at all before investing in topical breadth that this data says may not pay out there.
See your number
A free 60-second check shows what AI says about you.
Running this for clients? The $249 agency 5-pack audits five businesses, white-labeled.
Who runs this
- Built and operated by Sensara LLC, Atlanta, Georgia — about us and how the audit works.
- See what the report looks like before you run anything — score per engine, the competitors AI names instead of you, and a fix plan.
- We run the same audit on ourselves every week and publish the result: in the latest run AI named AskedAbout in 1 of 144 answers. We report our own numbers the way we report yours.