If you’re making content investment decisions based on how many topics you appear in across AI search, you might be optimizing for the wrong thing entirely.
Semrush recently published a study on how topical authority spreads in ChatGPT, and the findings should reshape how you think about AI visibility strategy. I want to walk through what matters most for CMOs trying to make smarter budget and content decisions.
Citations and brand mentions are not the same signal
This is the first thing to internalize. The study tracks two separate signals: citations (sources ChatGPT links to as evidence) and brand mentions (your brand name appearing directly in the response text).
They behave very differently, and conflating them leads to bad strategy.
Citations were relatively durable even in categories far from a brand’s core expertise. Brand mentions were not. In categories close to a brand’s core, 44% of appearances included a named brand mention. In distant categories, that dropped to 25%. And the combination of being both cited and named was nearly four times more common in close categories than distant ones.
That gap matters for you as a CMO because brand mentions are what drive recognition and recall. If ChatGPT is citing your domain as a source but not naming your brand, your customer never hears your name. They get the information you helped produce, but the attribution goes to the AI.
Know what you’re building toward. If the goal is brand recognition in AI responses, citations alone are not enough.
Depth beats breadth for brand mentions
This finding cuts against a lot of content marketing instincts. More topics, more coverage, more reach. That logic does not hold in AI search for brand mentions.
The study found that a brand appearing in only one of five prompts within a category was actually associated with a drop in mention share. The penalty didn’t disappear until the brand reached consistent presence across at least three of five prompts in that category.
For citations, breadth had less downside. Showing up once still generated some citation share. But for brand mentions, thin presence actually hurt.
The practical implication: if you’re spreading your content team across 30 topic areas trying to capture surface-level presence everywhere, you may be building citation share without ever building the kind of depth that gets your brand named. Categories where you consistently earn mentions are the ones from which you can credibly expand into adjacent territory.
Pick your anchor categories. Own them before you expand from them. This is prioritization as a discipline, not a mood: commit to the handful of categories worth owning this year, not the thirty that are theoretically true. A 30-topic content plan is not a plan; it is a way of postponing the decision.
Industry context changes the math
The study found that industry shapes how far topical authority travels, and this is important context for how aggressively to pursue breadth.
Finance and real estate showed more citation leverage from broader coverage. Showing up in all five prompts in a category was associated with stronger citation share, and spreading coverage across many categories didn’t appear to hurt performance.
Legal and healthcare were a different story. Even at full prompt coverage, brand mention share still trended negative. Being widely cited in those industries didn’t translate into being named.
If you’re in a regulated, high-trust industry, the bar to earn brand mentions in AI responses is higher. More content volume is not the answer. The signals that likely matter more are author credentials, third-party coverage from credible sources, primary research, and real evidence of expertise.
A practical framework for expansion decisions
The study’s recommendations translate cleanly into decisions you can make right now.
First, identify your actual anchor categories. These are not the categories where you’ve appeared once. They are the ones where your brand consistently earns both citations and named mentions across multiple prompts. If you don’t know which categories those are, that’s the first gap to close.
Second, when evaluating expansion targets, look for genuine proximity. The study measured closeness using semantic similarity, but in practice that usually means the customer, the problem, and the buying decision look similar to what you’re already answering well. Adjacency is not about topic keyword overlap. It’s about whether the underlying audience intent is related.
Third, do not confuse being cited with being recognized. Track both signals separately. If you’re building AI visibility for brand goals, you need to know whether ChatGPT is actually naming you or just pulling your content as an unnamed source.
The broader takeaway from this study is that AI search rewards depth and relevance over volume and reach. That’s actually good news for mid-market brands with focused expertise. You don’t need to cover everything. You need to genuinely own the categories that matter most to your audience and expand outward from a real position of strength.
Find out how to grow your marketing with the professional marketing minds at Ambient Array.
Original source: How topical authority spreads (and where it doesn’t) in ChatGPT