TL;DR Summary:
Category Framing: Your brand may be recognized in ChatGPT, but AI systems recommend brands based on the specific category language in the query, not just overall brand strength.Corpus Mismatch: If third-party content about your brand is tied to one category, it may surface there but disappear when users ask about an adjacent category with different language.Fix the Gap: The real solution is to expand the external content around your brand in the category your customers actually use, since Knowledge Graph updates alone do not change how AI recommends you.Why does my brand show up for one category but not another in ChatGPT?
You built a strong brand. You invested in PR. You cleaned up your schema markup and Knowledge Graph presence. But when people ask ChatGPT or Gemini for recommendations in your category, your brand doesn't appear.
The answer isn't about brand strength. It's about category framing and AI brand recommendations.
Recognition doesn't equal recommendation in AI models
Most brands assume AI visibility works like traditional search: build authority, and you'll rank for everything related to what you sell. That's not how large language models work.
LLMs don't evaluate your brand and decide if you're good enough to recommend. They evaluate the query and match it against category associations built from third-party content about your brand.
A seven-day study of 12 athletic apparel brands in the U.K. tested this directly. Researchers ran 14,140 API queries across ChatGPT, Gemini, Perplexity, Claude, and Google AI Overviews using the same brands but two different category terms: athleisure and athletic footwear.
The results showed that category framing and AI brand recommendations are not just connected. They're the same thing.
What happens when you change one word in the query
New Balance went from appearing in 1% of athleisure queries to 90% of athletic footwear queries. Lululemon went the opposite direction: from 90% in athleisure to 0% in footwear. The variation was roughly 90 points in both directions at the same time.
Nike, New Balance, and Reebok all share the exact same Google Knowledge Graph description: "Footwear company." All three are recognized perfectly by every LLM tested. But their behavior under different category framings was completely different.
Nike appeared in 77% of athleisure queries and 90% of footwear queries. New Balance appeared in 90% of footwear queries but only 1% of athleisure queries. Reebok landed somewhere in between.
This isn't correlation. This is a controlled observation showing what happens when you change only the category word in the prompt.
Category coding determines which queries surface your brand
The reason for the difference is what researchers call category coding: the combination of your Knowledge Graph description and the third-party content corpus that has accumulated around your brand in a given category.
The Knowledge Graph description anchors your brand to a category in the model's representation. This affects recognition.
The third-party corpus fills in what that category association looks like in practice. This affects recommendation.
New Balance's Knowledge Graph says "Footwear company." The third-party content about New Balance focuses on running shoes, performance footwear, and athletic training. When someone asks about athleisure brands, the model doesn't find New Balance in that corpus. It finds lululemon, Alo Yoga, and Gymshark because their corpus comes from fashion publications, lifestyle editorial, and activewear roundups.
When the query changed to athletic footwear, the retrieval flipped. New Balance was suddenly in the right corpus. Lululemon was not.
The model isn't making a judgment about brand quality. It's pattern-matching a query category against a content category. If those align, your brand surfaces. If they don't, it doesn't.
You can't fix category coding by updating your Knowledge Graph alone
Some brands will read this and think: change the Knowledge Graph description from "Footwear company" to "Apparel company" and the problem is solved.
That won't work.
The Knowledge Graph description is only half of what determines category coding. The other half is the third-party content corpus around your brand. That corpus doesn't change because you updated a field in the Knowledge Graph.
If your entire external content history focuses on performance footwear, running, and athletic training, changing the description gives the model a new anchor with nothing attached to it. The corpus still says what it always said.
The fix is third-party content investment in the specific category framing your customers use. You need coverage in the publications the model retrieves from, alongside the brands that already define that space. The Knowledge Graph description can support that work once the corpus exists.
How category framing and AI brand recommendations should change your strategy
Standard advice for AI visibility says to strengthen your entity: consistent name, clean schema, strong About page, more press coverage. That advice works for getting recognized and recommended within your coded category.
It doesn't work for getting recommended in adjacent category queries.
What determines recommendation in adjacent categories is whether the third-party content corpus around your brand matches the category framing your customers use.
Three questions matter:
Are you visible in AI?
What category has the LLM coded you into?
Is that the category your customers are querying?
If your brand is strong in one category but your customers are using adjacent category language to search, and your third-party corpus hasn't kept pace with that language shift, you'll be invisible in the queries customers are using.
Nike is the clearest positive case in the study. It surfaced in 77% of athleisure queries and 90% of athletic footwear queries, despite being Knowledge Graph-coded as a footwear brand.
The reason is that Nike accumulated enough athleisure-coded third-party content to register as category-eligible in both framings. It built a sub-stream in the adjacent category that New Balance didn't.
The diagnostic you should run before investing in entity optimization
Take the five or six different ways your customers might phrase a category query for what you do. Test each one across two or three LLMs. Note which formulations surface your brand and which don't.
For the ones that don't, ask:
Does third-party content about your brand use that language?
Are you written about in publications that cover that category?
Do you appear in editorial roundups that use that phrasing?
If the answer is no, you know where to start: getting into the external conversations that speak the language of that query.
To answer these questions systematically, you need to see what third-party content exists about your brand and in what context it appears. AI Mentions lets you track where and how your brand is being mentioned across the web, including the specific publications, the surrounding context, and which category language is being used alongside your brand name.
Instead of guessing whether you're being coded into "athleisure" or "athletic footwear," you can see the corpus that LLMs are drawing from: which articles mention you, what topics they associate you with, and which competing brands you're grouped alongside. This visibility makes it possible to identify corpus gaps before they become recommendation gaps and to track whether your third-party content strategy is shifting your category coding over time.
What closing the gap actually looks like
Closing that gap means becoming a participant in the category comparison content that defines who belongs in that space.
You need to appear in the same editorial contexts where brands already established in that category appear. You need journalists and editors to include you in roundups using the language your customers use. You need product reviews and comparisons that mention you alongside the brands the model already associates with that category.
This is not about creating your own content. Your owned content doesn't determine category framing and AI brand recommendations. The third-party corpus does.
You can't control what third parties write about you. But you can influence which conversations you're part of by placing yourself in the right editorial contexts and making it easy for writers to include you in the right category discussions.
The brands that win in AI aren't the ones with the strongest entities. They're the ones whose third-party corpus matches the language their customers use to search.
AI Mentions shows you which specific queries trigger competitor citations instead of yours, revealing the exact content gaps to fix. It tracks which competitor messaging AI assistants have absorbed that positions them as the default answer and tests whether your content updates improve AI mention frequency before you invest in full-scale production. If your brand is invisible in the category language your customers use, AI Mentions helps you diagnose the gap and prioritize the third-party content work that will close it.
