Semrush Data Shows AI Citations Cluster Around a Tiny Site Set
Semrush's 126-million-prompt AI Visibility Index shows citations clustering on a few sites per category — 65,000+ Patagonia mentions from six gear sites alone — and lays out a method to target them.

- Six outdoor-gear sites drove 65,000+ Patagonia mentions in AI answers, beating all Tier 1 media combined (Semrush AI Visibility Index).
- The Index analyzed 126 million prompts; a 2026 four-industry study found citations concentrate on a small site set on every AI platform tested.
- Finance core: NerdWallet, Investopedia, Bankrate out-cited Forbes and The Wall Street Journal.
- Recommended method: ~50 prompts across ChatGPT, Google AI Mode, Gemini and Perplexity; core = domains cited on 3+ of 4 platforms.
- Semrush advises re-running the citation core analysis quarterly.
Six specialist outdoor-gear sites generated more than 65,000 Patagonia mentions across AI answers — more than all Tier 1 traditional media combined, according to Semrush's AI Visibility Index. The figure is the clearest evidence yet for what Semrush now calls the "citation core": the small group of websites that AI platforms repeatedly cite when answering questions in a specific category.
The finding comes from the Index's analysis of 126 million prompts. A related 2026 study of AI answers across four industries found the same concentration pattern: a small set of websites earned most citations in each industry, on every AI platform tested.
Which sites dominate AI citations?
Semrush's category-level data shows the core shifts by vertical and geography. In U.S. finance, NerdWallet, Investopedia and Bankrate earned more AI citations than Forbes or The Wall Street Journal. In U.S. software, the core included G2, Capterra, TrustRadius and Forbes Advisor. Beauty citations clustered around Allure and Byrdie; travel around TripAdvisor.
Location matters too. AI answers about cars lean on Kelley Blue Book and Edmunds in the U.S., AutoTrader and What Car? in the U.K., and mobile.de and Auto Bild in Germany.
The method is illustrated with contract-management software vendor Juro, which appeared as a cited source for the prompt "best contract management software" on ChatGPT, Google AI Mode, Perplexity and Claude, and for related prompts such as "affordable contract management software."
Why does the citation core matter?
AI search treats third-party sources as a key signal for which brands to mention. Consistent brand presence across the sites AI draws on improves the odds of appearing in relevant answers, and citation core coverage can shape how AI presents a brand — favorably or not. These sites also influence buyers directly through reviews and comparisons.
How do you build your own citation core?
Semrush lays out a five-step process:
- Define scope. Pick a narrow category ("contract management software," not "software"), one target location, and the platforms to analyze — ChatGPT, Google AI Mode, Gemini and Perplexity are the recommended starting set.
- Build a prompt set of roughly 50 prompts spanning awareness ("What is contract management software?"), consideration ("Best contract management software for startups") and decision ("Juro vs. Ironclad") stages. Mine sales conversations, Search Console, and Semrush's Prompt Research tool.
- Collect citations. Record every cited domain and URL per prompt per platform — 50 prompts across four platforms means checking 200 responses. Semrush's Prompt Tracking tool automates this for ChatGPT, Google AI Mode and Gemini.
- Identify the core. For a four-platform analysis, keep domains cited on at least three. A pivot table or an AI tool can build the domain-level summary.
- Prioritize by topic relevance (High/Medium/Low), citation frequency, competitor gaps (sources citing rivals but not you, found via the AI Visibility Toolkit's "Missing" filter), and reachability — from directly editable profiles like G2 to low-influence targets like Wikipedia.
How do you earn placements on core sites?
The guide recommends six tactics:
- Claiming and updating review platform profiles (G2, Capterra, TrustRadius, Gartner Peer Insights, GetApp) and encouraging genuine reviews.
- Giving publishers original data — product usage, survey, business metrics or proprietary analysis — packaged with a methodology and pitched to relevant editors.
- Building research partnerships with academic sites; HubSpot's collaborations with Howard University and William Paterson University are cited as examples.
- Offering commercial partnerships to review and comparison sites, including affiliate commissions, referral fees, exclusive offers, or free product access.
- Participating credibly in communities like Reddit and Quora, mentioning products only when directly relevant.
- Using PR for major-media coverage and Wikipedia notability, treated as a long-term investment.
Does a placement guarantee AI mentions?
No. Semrush is explicit that citation core presence raises the chances of AI mentions — partly because web content feeds training data even without real-time search — but guarantees nothing. Progress tracking means re-running the same prompts monthly or quarterly, or using Prompt Tracking, which runs prompts daily and records changes in brand mentions, cited domains and mention rate.
The recommended cadence is a quarterly refresh, or sooner after a major market shift or changes to the AI platforms being tracked.
via docs.google.com (Original)
More from James Calloway
Show full bio
Market editor covering money, funding and market shifts around Search rankings and algorithm updates, generative engine optimization at SERP Journal.
231 articles