Should content strategy stay focused on a few core topics, or go broad to widen the funnel? New analysis by SEO analyst Kevin Indig, using Semrush AI Visibility Toolkit data, tests how topical authority behaves in AI search.
The headline finding: brands have good odds of being cited outside their established expertise, but not of being mentioned — that is, recommended.
The study uses Semrush's U.S. ChatGPT dataset covering 1,094 categories with five prompt variants per category, January through June 2026. The breadth models draw on 283,215 citations and 76,493 named-brand-mention observations, each paired with the following month's outcome. The relatedness test draws on 45,578 expansion appearances by 1,458 mapped brand entities. Every number is an association after controls, not proof of causation.
Embedding the ~1,000 categories and matching them against each brand's core expertise by semantic similarity produced a clear split.
In other words, you can be a source on almost any topic if you meet the normal criteria for citation. The AI recommends you in your core topics.
A domain can surface once in 20 categories and own none of them. Depth measures how many of a category's five prompt variants a domain appears in; breadth measures how many categories it appears in at all.
For citations, appearing across many categories carried no penalty. The association rises from +0.012 at 1-of-5 prompt appearances to +0.062 at 5-of-5.
Brand mentions behave differently. Appearing in only 1 of 5 prompts across the categories a brand shows up in is associated with a lower mention share at -0.051; at 5-of-5 it turns slightly positive. That is a small but robust effect: brands that actually own a category expand into others more successfully.
The "spread too thin" penalty is really a penalty for shallow presence, not for breadth itself. There is no defensible universal category limit, and the data does not support unlimited expansion either. The associations are also light — writing style, overall brand authority and third-party web mentions may outweigh topical authority entirely.
In finance, the citation-breadth association rises from +0.054 at one prompt variant to +0.139 at all five. Real estate rises from +0.021 to +0.135. In these markets, repeated broader coverage is associated with stronger source visibility, making category expansion a viable citation strategy.
In legal, the mention-breadth association stays negative even at full five-prompt coverage (-0.058); healthcare follows at -0.037. More consistent coverage improves citation position but does not reliably translate into named-brand visibility. These are higher-stakes decisions, where being a credible source is not enough to become the recommended answer.
Finance and real-estate teams can test adjacent-category expansion with citations as an early success metric. Legal and healthcare teams need a higher bar. A category is not won until the brand earns repeat mentions, not just citations.
For the mechanics behind concept-to-source association, see Entity SEO: AI Search Looks for the Source of a Concept, Not a Keyword; for the channel strategy that produces citations, see From Media Relations to Machine Relations.
A citation makes you a source; a mention names your brand in the answer. In categories distant from a brand's expertise, 50% of appearances were citations but only 25% were mentions.
Breadth itself does not — shallow presence does. Appearing in only 1 of 5 prompt variants per category is associated with a mention share of -0.051, while appearing in all five turns slightly positive.
Finance (+0.054→+0.139) and real estate (+0.021→+0.135) reward expansion with citation visibility, while legal (-0.058) and healthcare (-0.037) show no mention gain from broader coverage.
To apply what you just read to your own site, start with a free audit of where things are now.
A strategist replies within 24 hours on business days.
An E-E-A-T checker was never automatable because there is no API for it. Here is how an AI coding agent read Google's quality guidance, crawled a page sample, and produced a scored audit — plus where it stops.
A GEO audit is not checking whether your brand appears in an AI answer. Here is a nine-point GEO checklist split into four on-site items, four off-site, and one measurement layer.
LinkedIn, YouTube, Reddit, and TikTok are building systems to identify low-value AI content. Here are the limits of watermarking and what content still gets distributed.