AI Citation
An AI citation is when a generative engine, such as ChatGPT, Perplexity, or Google AI Overviews, names your page or brand as the source behind a statement in its answer, usually with a visible link. It is the generative-search equivalent of a search result, but the selection logic behind it is different: an AI decides, sentence by sentence, which sources back up which claims, rather than ranking pages against each other.
Why AI citations matter
People are increasingly getting answers directly from AI assistants instead of clicking through a results page. Pew Research Center found that when a Google search showed an AI summary, users clicked through to a traditional result in only 8% of visits, versus 15% when no summary appeared, roughly half the click rate. Ahrefs separately measured a 58% lower average click-through rate for the #1 organic result once AI Overviews appear for a query, and SparkToro/Similarweb put the zero-click share of US Google searches at 68% by early 2026, up from about 60% two years earlier. In that environment, being named inside the answer is often the only way you're seen at all: ranking well no longer guarantees a click, but a citation puts your name in front of the user regardless.
How an AI citation comes about
Many AI systems use retrieval-augmented generation: the model retrieves passages relevant to the question, then summarizes them and attaches sources to the claims it makes. Getting picked depends on two things clearing at once. The page has to be reachable by the crawler in question, and the content has to contain a passage the model can lift cleanly, a direct answer, a specific number, an unambiguous claim, rather than a paragraph of framing before the point. Ahrefs' analysis of ChatGPT citations found only 6-8% overlap with the same query's Google top 10, and about 80% of ChatGPT-cited URLs don't rank in Google's top 100 at all, so ranking well is neither necessary nor sufficient for being cited. Google has also said directly that AI Overviews and AI Mode require no special markup, schema, or AI-specific files to be eligible; the content still has to be genuinely useful, just not written differently "for AI."
Common mistakes
The most common one is chasing a technical fix instead of the content itself: publishing an llms.txt file and assuming it does something. It doesn't. Google's John Mueller has confirmed no Google Search system reads llms.txt, and an Ahrefs analysis of roughly 137,000 sites that published one found about 97% saw zero measurable referral traffic tied to it. A second mistake is writing pages the AI has nothing quotable to lift from: long throat-clearing intros, marketing language, no concrete claim to point to. A third is chasing backlinks as the main lever, when Ahrefs' study of roughly 75,000 brands found citation rate correlates with how often a brand is mentioned across the web at about 0.664, close to three times the correlation with backlinks alone at about 0.218. Third-party mentions, reviews, comparisons, forum threads, look like the stronger signal to build toward.
Relation to AI recommendations
An AI citation is factual: the assistant points to your page as evidence for a claim. An AI recommendation goes further and actively steers the user toward you, for instance in answer to "who's the best provider for X." The two build on each other: a brand that is cited often and in a favorable context is more likely to be treated as a credible option when the question turns evaluative. Citation rate, the share of relevant questions where you're named as a source, is a reasonable leading indicator of whether you'll eventually show up as the recommended pick, but the two are still distinct outcomes and should be tracked separately.
Example
A tax advisory firm in Leipzig publishes a page titled "How much does a simple tax return cost in Germany?" with a clear price range and what drives it up or down. Someone asks ChatGPT the same question, and the assistant answers with a number range and names the firm's page as its source, with a link. That's an AI citation: the firm shows up inside the answer even though the person never opened a search engine or saw a results page at all.
Common questions
How is an AI citation different from a Google ranking?
A ranking is a position in a list the user browses and clicks through themselves. An AI citation is your page being named as the source inside a finished AI answer. Ahrefs found only 6-8% overlap between URLs ChatGPT cites and the same query's Google top 10, so the two are largely separate selection processes, not the same competition measured two ways.
Does adding schema markup or an llms.txt file get me more AI citations?
Not on the evidence available. Google states no special schema is required for AI Overviews or AI Mode, and Google's Mueller has confirmed no Google Search system reads llms.txt at all; an Ahrefs study of about 137,000 sites that published one found roughly 97% got no measurable referral traffic from it. The stronger lever, per Ahrefs' analysis of about 75,000 brands, is being mentioned across the web more often, which correlates with citation rate nearly three times as strongly as backlinks do.