LLM Citation

When an AI assistant answers a question, it typically draws on a mix of its trained knowledge and, for tools that browse or use retrieval (see RAG), a set of documents fetched at answer time. A citation is what happens when the model attributes a fact or recommendation to one of those documents — often shown as a footnote, a linked source, or a named business inside the prose of the answer.

Being cited is now a meaningful source of enquiries for small and regional businesses, because a citation functions like a recommendation rather than a listing: the assistant is telling the user which business it trusts enough to name. Businesses earn citations by publishing content that is specific, factual, dated and easy for a machine to parse — the same discipline that underpins generative engine optimisation.

Citations cannot be bought or guaranteed the way paid search placement can. They are earned through structure, clarity and consistency, and they can change from one query to the next as models are updated and re-crawl the web.