Login
What co-mentions reveal about the AI recommendation gap
We’ve spent the last two years optimizing for AI visibility by focusing on what we say about ourselves: writing better About pages, adding clear schema and SameAs markup, structuring content more effectively, and providing more direct answers.
All of these principles still apply and are essential for the qualification phase of an LLM’s brand processing (clarity + relevance). But a study João da Silva and I conducted using Friction AI’s platform puts a number on a factor the industry has been circling around but couldn’t prove.
Among brands that were already recognized (where the LLM could describe them accurately), Knowledge Graph (KG) strength predicted visibility within the category each brand was coded to. What it didn’t predict was whether a brand would surface in an adjacent category query, even if it belonged there from a business perspective. In other words, recognition didn’t guarantee recommendation. That’s the framing gap.
What brands did we test and how did we test them?
For this case study, we tested 12 athleisure and activewear brands across ChatGPT, Gemini, Perplexity, Claude, and Google AI Overviews: 14,140 API runs over seven days, using UK geography with web search enabled.
For each brand, we ran two types of prompts:
- Recognition prompts (“What is [Brand]?” and “Describe [Brand]”)
- Recommendation prompts (“Best athleisure brands,” “Top 10 athleisure brands,” and “Which athletic apparel brands are worth buying in 2026?”)
The brands spanned three Knowledge Graph tiers, assigned by Google KG resultScore (the raw score returned by Google’s Knowledge Graph Search API — a proxy for how strongly an entity is established in Google’s index), so we could test whether KG strength predicted recommendation visibility:
- Low KG: LNDR, TALA, Gymshark, Varley.
- Mid KG: Reebok, Outdoor Voices, Rhone Apparel, Sweaty Betty.
- High KG: Alo Yoga, Nike, lululemon, New Balance.
Spoiler ahead: The high-KG brands didn’t dominate recommendations. The mid-KG tier showed the largest average gap between recognition and recommendation.
Within the high-KG tier, some brands were universally recommended, while others were nearly invisible in recommendation prompts, despite being perfectly recognized across every LLM we tested.
The SEO toolkit you know, plus the AI visibility data you need.
Recommendations: What the co-mention data showed
We mapped how often brands appeared together in athleisure content across external sources (articles, reviews, comparison pieces, and editorial lists) crawled via API from UK-indexed sources.
Some of the most interesting results include:
- lululemon + Alo Yoga: 534 co-mentions.
- lululemon + Nike: 482 co-mentions.
- Alo Yoga + Nike: 449 co-mentions.
- Gymshark + lululemon: 264 co-mentions.
- Gymshark + Alo Yoga: 252 co-mentions.
These brands appear together repeatedly in the same articles, roundups, and editorial comparisons across independent sources. Together, they form a cluster that the LLM treats as “athleisure.”
Now look at the other end of the spectrum. New Balance co-occurs with lululemon in athleisure content so rarely that it doesn’t appear in the top pairs at all. Nike co-occurs with lululemon roughly 50 times more often than New Balance does.
Nike, New Balance, and Reebok share the exact same Google Knowledge Graph description: “Footwear company.” From an entity standpoint, they start from the same position. But Nike is inside the athleisure cluster. New Balance and Reebok are entirely outside it.
The LLM isn’t evaluating these brands independently and deciding which ones fit athleisure. It’s pattern-matching against associations built from external content. If a brand hasn’t appeared consistently alongside lululemon, Alo Yoga, and Gymshark in the content the model trained on — or retrieves from — it doesn’t belong in that cluster because the semantic association was never built.
Nike, the hero: Same KG description, completely different results
Nike, New Balance, and Reebok share the same KG entity description: “Footwear company.” LLM probing across all five systems assigns all three unanimously to the athletic_footwear category, so from a pure entity-clarity standpoint, they start from the same position.
However, their recommendation rates in athleisure queries aren’t remotely equivalent.
Nike surfaces in 71% of athleisure recommendation prompts, while New Balance and Reebok appear in 0% across all five LLMs and all 14,140 runs.
The difference isn’t how they’re defined (“Footwear company”). It’s which conversations they appear in and which other brands appear alongside them.
LLMs don’t infer category adjacency. If a brand hasn’t been consistently mentioned alongside the relevant players in a category — in press, reviews, editorial content, and comparison pieces — the model doesn’t make the leap. Jason Barnard describes this well: if A plus B should equal J, you have to construct that path explicitly. The model won’t build it for you.
New Balance’s co-mention density lives in running and performance content. Nobody built the semantic bridge from running → athletic lifestyle → athleisure in external content, so the model doesn’t cross it. The Knowledge Graph says “Footwear company,” and the third-party corpus confirms footwear. Athleisure queries retrieve the athleisure corpus, and New Balance isn’t in it.
The third-party citation weight in recommendation vs. recognition data
When we split citations by prompt type — recognition vs. recommendation — a pattern emerges that should reframe where most GEO budgets are being spent.
For recognition prompts — where the user has already typed your brand name — own-brand content is the dominant source:
- ChatGPT cited own-brand content 49% of the time.
- Perplexity: 36%.
- Claude: 23%.
This is where your About page and homepage are used for clarity, and your services, category, and guide pages are used for relevance.
Recommendation prompts give us completely different results. When the user hasn’t named your brand and is asking for the best option in a category, own-brand citations drop to 18% on ChatGPT and to effectively zero on Gemini, Claude, Perplexity, and Google AI Overviews. Third-party sources account for 82% to 100% of what gets cited across all five systems.
The GEO community has argued for some time that external signals matter more than on-site optimization for recommendation visibility, and this data puts specific numbers behind that argument. It also shows that external signals aren’t all the same thing.
- Entity clarity gets a brand recognized. That’s a problem you solve on your own site.
- External credibility gets it considered. That’s a PR and corroboration problem.
- Co-mention density in the right category cluster places a brand in the concept graph for a specific recommendation query. That’s a category-positioning problem.
These are three separate problems that require different solutions. Conflating them is why many GEO recommendations stop short.
The practical addition to any GEO audit is this: after checking entity clarity and external credibility, audit where you appear in relation to others.
- Are your press mentions listing you alongside your actual category competitors?
- Do the roundups that include you also name the brands that dominate your target category?
If not, the LLM has probably never learned to associate you with that category because it has never seen you in that “company.” Unlike entity clarity or schema, it’s not something you can fix on your own website. That’s the gap.
What the co-mention structure means for PR and content strategy
As we’ve seen so far, being mentioned in a category isn’t enough. Being mentioned alongside the right brands in a category is what places you in the concept graph for that cluster.
A press mention that describes a brand as “performance apparel” in isolation does little to advance its athleisure concept graph placement.
A press mention that lists it alongside lululemon, Alo Yoga, and Gymshark in an editorial comparison does considerably more because it builds the co-occurrence signal the model needs to associate the brand with that cluster.
The same logic applies across content type.
Editorial roundups and comparison pieces
Being included in “best of” lists that name your category competitors is worth more to your concept graph than a standalone brand profile. The cluster signal comes from appearing in the same article as the brands that define the category.
Podcast appearances
If the host introduces you in relation to specific named brands, or compares your approach to a category leader, that co-occurrence gets indexed.
A bio that says “founder of [Brand], which competes with lululemon and Gymshark in the premium athleisure space” does different work than a bio that says “founder of [Brand], a performance apparel company.”
Analyst and industry reports
Category-level reports that group brands together are high-signal co-mention sources. Being included in a sector analysis alongside your category peers places you in the concept graph in a way that standalone coverage doesn’t.
Retailer and comparison taxonomy
Being stocked and categorized alongside category leaders in a major retailer’s taxonomy is a co-mention signal. The retailer’s category page is external content that places your brand in a cluster.
The goal is visibility in the right company.
Track, optimize, and win in Google and AI search from one platform.
A note on the data and what comes next
This study covers a single category — athleisure and activewear — with 12 brands tested in the UK. The co-mention figures are raw co-occurrence counts from UK-indexed sources crawled via API, covering content indexed at the time of the study in May 2026. Cross-category validation and additional geography testing are in progress.
The full paper, “The Recognition-Recommendation Gap: Empirical Evidence That Category Coding, Not Knowledge-Graph Strength, Determines Brand Visibility in Generative AI Output,” has been published by João da Silva and me on Zenodo and documents the methodology, brand sample, prompt set, and extraction code in sufficient detail for independent replication.
But the pattern in the co-mention data is clear enough to act on now. Three brands share the same Knowledge Graph description: one appears in 71% of athleisure recommendation responses, and two appear in 0%. The structural difference is co-mention density in category-aligned third-party content.
The question worth asking about any brand is this: In the content that talks about your category, are you in the room, and are you in the right company?
