Definition
An AI citation is a source reference that an AI answer engine displays or relies on when it generates an answer to a user’s question. It links a specific claim in the generated answer back to a web page, document, or other source. For answer engine optimization teams, citation tracking shows whether answers lean on owned pages, third-party pages, competitor pages, or no visible source at all.
Citation vs Mention
A citation is not the same as a mention. A mention names a brand or product inside the answer text. A citation attaches a source URL or reference to a claim. A brand can be mentioned without being cited, cited without being mentioned by name, or both at once. Tracking the two separately matters because they map to different fixes: mentions reflect how the model already understands a brand, while citations reflect which crawlable pages it treats as evidence. See the related concept of an AI mention for the naming side of this distinction.
Why AI Citations Matter
Citations reveal which pages AI systems treat as trustworthy evidence for a topic. When an answer engine cites a page, it is effectively recommending that source to the user, often inside a zero-click experience where the buyer may never run a traditional search. If competitor pages earn the citations and yours do not, the brand can be invisible in the exact moment a decision is forming, even when classic rankings look healthy. This is why citation coverage is a core signal in AI visibility reporting and why it sits alongside rankings rather than replacing them. The shift it represents is covered in depth in our explainer on AEO versus traditional SEO.
How AI Engines Select Sources To Cite
Answer engines do not cite from a fixed ranking. They retrieve candidate passages, evaluate relevance to the prompt, and attach sources to the claims they synthesize. Pages tend to be cited when they are crawlable by the relevant AI user agents, return stable responses, answer the question directly near the top, name entities consistently, and expose original evidence a model cannot infer from common web consensus. Login-gated, JavaScript-only, or contradictory pages are harder to retrieve and attribute.
How To Improve Citation Coverage
Improve citation coverage by making source pages crawlable, direct, well structured, internally linked, and supported by visible proof such as definitions, methodology, screenshots, or first-party data. Schema should match visible content; it should never invent claims the page does not support. Consolidate thin variants into one stronger canonical page rather than spinning a separate page per query, then re-test the prompts that matter to confirm the right URL is being selected. The repeatable workflow for this is described in our AI citation tracking methodology, and the software side is covered on AI citation tracking.
Related Terms
Related terms include AI citation tracking, cited source URLs, source coverage, answer engine optimization, share of mind, and AI visibility.