Topical authority in ChatGPT does not spread evenly. A new Semrush study published in 2026, analyzing 283,215 domain-category citation observations across 1,094 US categories tracked from January through June 2026, draws a sharp line between two outcomes brands often conflate: being cited as a source and being named as a brand. They follow different rules, and mixing up the strategies for each is where most content programs lose ground.
Are Citations and Brand Mentions the Same Signal in ChatGPT?
No, and the gap between them is larger than most brands expect. In the Semrush dataset (January through May 2026), citations (the sources ChatGPT links to as supporting evidence in its answers) and brand mentions (the brand names ChatGPT includes directly in its response text) respond to different inputs. Citations travel further from a brand’s core expertise. Brand mentions stay much closer to home.
The numbers are specific. In categories close to a brand’s core expertise, Semrush data shows brands were cited in 74% of appearances, mentioned by name in 44%, and both cited and mentioned in 34%. In categories distant from core expertise, those rates fell to 50% cited, 25% mentioned, and just 9% both.
The cited-and-named rate was nearly 4x higher in close categories than in distant ones. And of all citations that occurred in the least related categories, only 18% also included a named brand mention. In the most relevant categories, 46% did. According to Semrush, related categories were more than twice as likely to turn a citation into a named-brand mention in this sample.
Does Depth in One Category Beat Spreading Across Many?
For citations, spreading across categories carries no obvious penalty. Semrush data (January through May 2026) shows a brand appearing in even one of five prompt variants (five slightly different phrasings of the same category question) in a category tended to gain some citation share. No “spread-thin effect” was observed for citations alone.
Brand mentions follow a different rule. A brand appearing in only one of a category’s five prompt variants was associated with a drop in brand-mention share. That penalty only disappeared once the brand reached at least three of five prompts consistently. Showing up once across 20 categories did not build higher brand-mention share in this dataset.
The study puts it directly: “The categories where a brand consistently earns named mentions are the ones from which it can credibly expand.”
Does Industry Shape How Far Topical Authority Travels?
Yes, meaningfully. In finance and real estate, the Semrush data (January through May 2026) shows wider coverage held more citation leverage. Brands appearing in all five prompt variants in a category were associated with stronger citation share than brands appearing in just one. Spreading across categories in these industries did not appear to hurt citation presence.
Legal and healthcare set a higher bar. Citation gains from broader coverage were weaker than in finance or real estate. More notably, even full five-prompt coverage in legal and healthcare still trended negative for brand-mention share. Being widely cited in those fields did not translate into being named.
What This Means for AI-Search Visibility
The clearest takeaway from this data: building citation presence and building brand-name recognition in ChatGPT are two separate goals that require different investments, not one content strategy that delivers both at once.
Think of it this way. If a financial publication gets cited in a ChatGPT answer about homeowner’s insurance, that is a citation. ChatGPT naming that publication as a recommended brand for homeowner’s insurance questions is a much higher bar. The first is about content quality and relevance matching. The second is about how consistently and deeply the model associates that brand with a specific category across many different queries.
One pattern the study data implies but does not state directly: the ceiling on brand-mention expansion in legal and healthcare is likely not a content-volume problem. If full prompt coverage still trends negative for mention share in those industries, the bottleneck probably sits in off-content signals: institutional credibility, author qualifications, and the kind of third-party coverage (from professional associations, peer-reviewed sources, or established trade media) that AI systems treat as authority signals in high-stakes domains. Publishing more pages is unlikely to move that needle on its own.
The Semrush data also suggests a reframe for how brands measure AI-search progress. Tracking only citation counts can mask a brand-mention deficit. A brand could increase its citations by 30% while its named-mention share stays flat or falls. Separating the two metrics is not just methodologically useful. It reveals whether a content program is building the kind of presence that influences customer decisions or simply feeding AI reference pools without accruing brand recognition.
Hingewise’s read: for most brands, topical authority is a depth-first problem before it is a breadth problem. The data supports expanding outward from a small set of categories where ChatGPT already names you consistently, not from categories where you have appeared once or twice. In regulated industries, the entry point is different and starts with credibility signals rather than content volume. That is a harder and slower path, and brands in those industries should plan for it accordingly rather than benchmarking against finance or e-commerce timelines.
Before Expanding into a New AI Search Category: A Quick Check
- Confirm you appear in at least 3 of 5 prompt variants in your current anchor categories, not just occasional appearances.
- Check whether you are being cited only or cited and named. The gap between those two rates signals how ChatGPT currently positions your brand in that category.
- Identify target categories by how closely they match the customer, problem, and buying decision in your anchor categories, not just by topic adjacency.
- Audit third-party coverage in the target category: industry publications, review sites, and expert commentary that associates your brand with that space.
- Review author credentials and original research on your existing content in the target category before adding more pages.
- In legal or healthcare contexts, assess whether broader citation gains are actually producing brand-mention growth before scaling content investment further.
- Track citations and brand mentions as separate metrics. Improving one does not automatically improve the other.
What to Watch Next
The Semrush and Growth Memo dataset covers US categories on ChatGPT from January through June 2026. Whether the citation-to-mention gap observed here holds across other AI search platforms such as Perplexity, Gemini, or Google AI Overviews (AI-generated summaries that appear at the top of Google search results), and how it shifts as models update, remains an open question. Industry-level patterns in regulated fields are particularly worth monitoring as AI platforms develop vertical-specific trust and authority signals.
Source: Semrush Blog, “How topical authority spreads (and where it doesn’t) in ChatGPT,” 2026. Data partnership: Kevin Indig / Growth Memo. Dataset: Semrush AI Visibility Toolkit.
Lam Nguyen · Hingewise