AI citations are the new impressions
Buyers now ask an assistant before they ask a search engine. Here is how to measure whether ChatGPT, Perplexity, Gemini and AI Overviews are citing you, and what to do when they are not.
Key takeaways
- Citation share is measurable per surface and per prompt cluster.
- Answer-first pages with a definition and a Q&A block are quoted far more often.
- A press page with structured data is one of the highest-leverage assets you can publish.
Search used to be a list of links. It is becoming an answer. When a buyer asks an assistant what tool to use for a job, the assistant does not return ten results — it returns one recommendation with a handful of citations. If you are not in that handful, you were not in the conversation.
That changes what public relations is for. A press page is no longer just a folder for journalists. It is a corpus that assistants read, quote and cite. The same is true of a well-structured release, a guide that answers a question directly, and a comparison page with facts rather than adjectives.
We measure citation share across four surfaces: ChatGPT, Perplexity, Gemini and Google AI Overviews. For each surface we track a cluster of prompts a buyer in your category would actually type, then record whether your brand appears, how prominently, and which page was cited.
The pattern across the accounts we run is consistent. Pages that open with a one-sentence definition get quoted. Pages that bury the answer under three paragraphs of context do not. Pages with FAQ structured data get quoted more often than identical pages without it. Pages with a source link next to each claim are quoted with more of the claim intact.
The practical implication is that the work of earned media and the work of answer engine optimisation are now the same work. Earn the coverage, publish the facts, structure the page, and keep the sources visible. That is the whole strategy, and it compounds.
Written by PRMagi Research.