AI citations are the source links AI answers attribute information to. Learn how AI engines decide what to cite, and how to earn citations.
Frequently asked questions
What is an AI citation?
An AI citation is a source link or reference that an AI answer attributes information to. When ChatGPT, Perplexity, or Google's AI Overviews cite your page, they credit it as a source and can send the reader to your site.
How do I get cited by ChatGPT?
Get cited by ChatGPT by publishing authoritative, well-structured content that answers specific questions directly, earning mentions on sources it trusts, adding schema, and keeping facts current. Strong, retrievable content that clearly answers the prompt is the foundation.
Why isn't my site cited in AI answers?
Common reasons include thin or unstructured content, weak topical authority, no presence on third-party sources the engine trusts, missing schema, or stale information. Often the page simply doesn't answer the specific question in an extractable way.
Do backlinks affect AI citations?
Indirectly, yes. Backlinks and reputable mentions build the authority and corroboration that make an engine more likely to trust and cite your content, but clear structure, relevance, and accurate, current information matter just as much.
An AI citation is the source link or reference an AI answer attributes its information to. When ChatGPT, Perplexity, or Google's AI Overviews cite your page, they credit it as a source, and can send the reader to your site. In AI search, citations are the unit of visibility: where traditional SEO counts clicks from a ranking, AI search counts whether the answer names and links to you.
Key takeaways
An AI citation is a source an AI answer credits, and a path back to your site.
Engines cite by authority, relevance, structure, freshness, and source trust.
You earn citations with extractable content, entity authority, and trusted mentions.
Backlinks help indirectly, by building the trust that makes a citation likely.
Citation share is measurable, and the clearest signal that your AEO is working.
Most answer engines retrieve candidate sources, generate an answer grounded in them, and cite some of those sources. Which ones get cited comes down to a consistent set of factors.
It starts with relevance. The source has to match the meaning of the question, not just its keywords, so content that addresses the specific prompt directly is easier to retrieve and quote.
Then there's authority and source trust. Engines favor sources they have reason to trust: established sites, pages corroborated by mentions elsewhere, brands with a consistent, verifiable identity. Off-site reputation feeds on-answer citations.
Structure matters too. A model can pull a clean, accurate snippet from a clear heading, a short paragraph, a list, or a table far more reliably than from dense prose. Extractable structure makes your content safe to cite.
And freshness counts. For anything time-sensitive, engines prefer recently updated sources, and stale facts get passed over, or repeated incorrectly.
You can't edit the answer or see the index, but you can be the most relevant, trusted, well-structured, and current option for the questions that matter to your brand. The same factors drive whether you appear in Google AI Overviews and every other engine.
Earning citations is a discipline, not a hack. The moves with the most leverage:
Publish citable, structured content. Lead with a direct, self-contained answer to a specific question, then back it with clear headings, lists, and tables. The easier a passage is to lift accurately, the likelier it is to be cited.
Build entity authority. Be consistent and unambiguous about who you are across your site and the web, and reinforce it with Organization, Article, and FAQPage schema so engines can resolve and trust your brand.
Earn mentions on sources engines trust. Accurate references in roundups, reviews, reputable publications, and documentation raise the corroboration that makes a model confident citing you. It's the citation version of link building.
Use structured data. Schema removes ambiguity about what your content is, which helps engines parse and reuse it correctly.
Keep facts current. Date pages, update figures, and prune stale claims. Freshness is both a retrieval signal and an accuracy safeguard.
Done together, these make your content the kind of source an answer engine reaches for, and they're the core of answer engine optimization.
Earning citations is only half the job. You also need to know whether it's working. Because AI answers are personalized and short-lived, you measure citation share by sampling: run the prompts that matter, across the engines your audience uses, and record how often each answer cites your site versus competitors.
Track citation share alongside plain mentions: a mention means the answer named you, while a citation means it linked to you, and the gap between the two is itself a signal. Doing this by hand doesn't scale, so brands use AI visibility tools to monitor citations on a schedule. Elmo is an open-source tool that tracks which AI answers cite your brand across the major engines, and our guide to tracking your brand in AI search lays out the full method.
A few recurring mistakes keep otherwise-good brands out of AI answers:
Thin content. Pages that restate a query without genuinely answering it give an engine nothing worth quoting.
Unstructured pages. A wall of prose with no clear headings, lists, or direct answers is hard to extract from, so it gets skipped.
No schema or entity clarity. Without structured data and consistent identity, engines struggle to resolve who you are and what you're authoritative about.
Stale information. Outdated facts get passed over for fresher sources, or, worse, repeated wrongly.
Chasing citations without authority. Formatting tricks won't earn citations if nothing off-site corroborates your brand. Reputation and structure work together.
Avoiding these is mostly a matter of publishing genuinely useful, well-structured, current content and building the authority to back it. For the vocabulary behind these concepts, see the AI search glossary; for the wider tooling, the AI visibility software hub.