What Are LLM Citations?
LLM citations are the sources an AI engine references when it answers a question. Definition, why they decide which brands get recommended, and how to win them.
LLM citations are the sources an AI engine references when it answers a question: the specific web pages, articles and documents a large language model draws on and links to as the basis for its response. On engines that show their work, the citations appear as numbered links or source cards next to the answer, telling the reader where each claim came from. They are the receipts behind an AI recommendation.
A citation is not quite the same as a mention. A mention is the engine naming your brand in the text of an answer; a citation is the engine pointing to a page as a source. You can be cited without being named, and named without being cited, but the two reinforce each other, because the pages an engine cites are the raw material it uses to decide which brands to name in the first place.
- An LLM citation is a source an AI engine links to as the basis for part of its generated answer.
- Perplexity and Google AI Overviews show citations openly; ChatGPT, Gemini and Claude cite too, though how visibly depends on the mode.
- Citations shape recommendations: the pages an engine trusts are the pages that decide which brands it names.
- To earn citations, be present in the sources engines already trust and structure your content so answers are easy to extract.
Which engines show citations
Perplexity is the clearest example. Built as an answer engine, it attaches a numbered list of sources to almost every response and surfaces them prominently, which makes it the easiest place to see your citation footprint at a glance. Google AI Overviews, the AI answer that now sits above the classic blue links on many searches, also show citations as links inside and beside the generated summary, so you can trace which pages fed the answer.
ChatGPT, Gemini and Claude cite as well, though how openly depends on how they answer. When these assistants browse the live web, they typically list the pages they used; when they answer from training data alone, there may be no visible source at all. Measure tracks citations and mentions across all five of these engines, so you can see which sources win across the whole set rather than checking one at a time.
Why citations decide who gets recommended
Citations are not decoration. They are the evidence a model assembles its answer from, so the set of pages an engine cites for a question is effectively the shortlist it reasons over. If your category's answers keep citing the same three review sites, two competitors' documentation pages and a community thread, those sources are shaping which brands get recommended, and a brand absent from all of them is unlikely to be named.
This is why citation analysis is the practical core of AI search visibility. Tracking whether you are mentioned tells you the outcome; tracking which sources are cited tells you why, and where to act. Change the sources and you change the outcome.
A mention is your brand named in the answer text. A citation is a page the engine links to as a source. Mentions tell you the score; citations tell you the reason behind it.
How to get cited by AI engines
Getting cited comes down to two things: being present in sources the engines already trust, and making your content easy for a model to lift.
- Earn a presence in trusted sources. Engines lean on established, high-authority pages: reputable publications, well-regarded review and comparison sites, documentation and active community threads. Being reviewed, listed and discussed on those pages matters as much as your own site, because that is where models look first.
- Structure content for extraction. Lead with a direct answer, use clear headings and keep claims specific and self-contained. Definitions, comparison tables and short factual statements are easy for a model to quote verbatim, which makes them more likely to be cited.
- Keep facts consistent and current. Maintain accurate, up-to-date pricing, product details and claims across the web. Contradictory or stale information makes a source less reliable, and therefore less likely, to cite.
- Stay accessible to AI crawlers. If engines cannot fetch your pages, they cannot cite them. Confirm your robots rules and llms.txt allow the crawlers you want with a free AI crawler checker.
Optimizing deliberately for all of this is the discipline of answer engine optimization (AEO), which works backward from how engines pick what to cite.
How to track your LLM citations
You cannot improve what you cannot see. Track a consistent set of your buyers' real questions across the engines, record which sources each answer cites, and watch the trend over time. Perplexity is the best place to start because its citations are so visible; our guide to tracking AI visibility in Perplexity walks through it step by step.
Measure does this across ChatGPT, Perplexity, Gemini, Claude and Google AI Overviews, records the exact sources winning each citation, names the competitors cited instead of you, and ties what it finds to the GA4 traffic those answers drive. You can start tracking your citations free or explore the free AI visibility tools first.
Frequently asked questions
What are LLM citations?
LLM citations are the web sources an AI engine references when generating an answer. Engines like Perplexity and Google AI Overviews show them explicitly. They matter because the brands mentioned in cited sources are the brands the answer tends to recommend.
How do I get cited by AI engines?
Be present and clearly described in the sources engines already trust for your topic: independent roundups, reviews, comparison pages and your own structured content. Tracking which domains win citations for your questions shows exactly which pages to target.
We build Measure, a self-serve platform that tracks how AI engines describe and recommend brands, and connects that visibility to real traffic and revenue. Everything here is written from what we see in the data every day.