What is an AI citation?
An AI citation is a reference to a brand, page, or source inside an AI-generated answer, whether in Google AI Overviews, ChatGPT search, Perplexity, or Claude. Citations are the unit of visibility in answer engines: where classic SEO counts rankings and clicks, AEO counts how often and how favourably you are cited.
How AI engines choose what to cite
Different engines select citations differently. Perplexity and ChatGPT search retrieve live web results and cite the pages they synthesize from, which makes retrieval-friendly content (clear headings, direct answers, factual density) the entry ticket. Model-internal answers without live retrieval lean on training data, where Wikipedia presence, consistent entity descriptions, and authoritative coverage determine whether a brand gets mentioned at all.
Citation tracking is becoming a standard reporting line: query a fixed prompt set across engines on a schedule, record which domains get cited, and measure share against competitors. The pattern that wins citations consistently is the citation capsule: a self-contained passage stating the claim, the evidence, and the source in a form an LLM can lift cleanly.
Example
Example
Asked for "best portable storage container companies", Perplexity cites two industry roundups and one vendor page. The vendor earned its citation with a comparison table and plain-language pricing: content structured for retrieval, not keyword density.
Frequently asked questions
How do AI engines choose what to cite?
Retrieval-based engines cite the pages whose passages best answer the query, favouring direct, factual, well-structured content. Engines answering from model memory mention brands with strong, consistent entity footprints across authoritative sources.
Can you measure AI citations?
Yes. Run a fixed set of buyer-relevant prompts across engines on a recurring schedule and log cited domains. Tooling now automates this, producing a share-of-citations metric comparable to rank tracking.