AI citation
The core unit of AI visibility: when an answer engine references your domain as a source for what it just said.
An AI citation is an instance where an answer engine — ChatGPT, Gemini, Grok, Perplexity or Google’s AI Overviews — references your domain as a source for an answer it generates, typically as a linked or named attribution. It is the core unit of AI visibility: the thing you count to know whether you exist in the answer layer at all. A citation differs from a plain mention: a mention names your brand in passing (“tools like Acme and Beta”), while a citation attributes a claim or recommendation to you as the source. Both are valuable, but a citation is the stronger signal because it means the engine treated your page as evidence. The full set of queries you are cited on is your AI citation footprint, and the proportion of relevant answers that cite you is your AI share of voice.
What counts as an AI citation?
At minimum, an AI citation is the engine attributing part of its answer to your domain. In practice it shows up in a few forms:
- Linked citation — the answer includes a clickable source link to your URL (common in Perplexity, AI Overviews, and ChatGPT’s browsing mode).
- Named source — the answer names your brand or site as where the information comes from, with or without a live link.
- Recommendation — the answer recommends your product or service by name in response to a buying-intent question.
Citation vs mention: what’s the difference?
A mention is your name appearing in the answer text; a citation is your domain being credited as a source. The distinction matters because they are acted on differently: mentions tell you the model associates your brand with a topic, while citations tell you a specific page earned the attribution and can be reinforced. We unpack the practical implications in mention vs citation in AI.
Why are AI citations the metric that matters?
Because in AI search the answer often replaces the click — a zero-click outcome — the citation is the visibility. If the engine doesn’t name you, you were effectively absent from that buyer’s decision, no matter how well you rank in classic search. Counting citations, query by query, is therefore the honest scoreboard for AEO, GEO and LLMO alike.
How do you find your AI citations?
You find them by reading the AI query index backwards — starting from your domain and listing every query it is cited on. That is reverse AI search. The free Domain Check does it instantly: enter a domain and see the real ChatGPT, Gemini and Grok answers that cite it, plus the competitors cited alongside — no signup, no score, just the citations.
Worked example
A user asks an assistant “what’s a good budget standing desk?” The answer reads: “A frequently recommended budget option is the FlexiDesk (flexidesk.example), praised for its stability.” That is an AI citation: flexidesk.example is credited as the source for the recommendation. If the answer instead said “brands like FlexiDesk and others make budget desks” without attributing anything, that would be a mention, not a citation — the brand is named, but no claim is sourced to it.
Related terms
- AI citation footprint — the full set of queries you are cited on.
- AI share of voice — your citations as a share against competitors.
- Zero-click — why a citation can matter even with no visit.
Frequently asked questions
What is the difference between an AI citation and a mention?
A mention names your brand in passing; a citation credits your domain as the source of a claim or recommendation. A citation is the stronger signal because it means the engine treated your page as evidence.
Why are AI citations the metric that matters?
Because in a zero-click world the answer often replaces the visit, so the citation itself is the visibility. If the engine does not name you, you were effectively absent from that buyer’s decision.