Not every platform cites in the same way. Perplexity was built around cited answers: its responses typically carry inline citations to retrieved sources, and a reader can follow them back to the original pages. ChatGPT cites sources in its web-connected modes, though the number and placement of citations vary by query. Google AI Overviews present a summarized answer above traditional search results, with links to the pages that informed the summary. Gemini similarly links to sources in many of its answers, with the depth of attribution depending on the question and the mode. The details change as these products evolve, which is why this section treats platform behavior as something to observe and document rather than as settled fact.

Two distinctions matter. The first is between a citation and a mention. A business can be mentioned in an AI answer — named in the text — without being cited as a source, and it can be cited without the answer’s main claims resting on its pages. Citation share is a way of measuring this: across a set of questions, how often does a given source appear among the cited sources, compared with its competitors? It is a useful diagnostic, but it is not a ranking and not a measure of quality on its own. The second distinction is between citation and endorsement. An AI system may cite a page that criticizes a company, a page that merely lists it, or a page whose claims it has garbled. Being cited means being used as material, not being recommended.

For businesses, the practical question is what makes a source citable. The patterns are consistent with how retrieval works: sources that are accessible, clearly structured, and corroborated by other independent sources are easier for a system to retrieve and to cite with confidence. A single, isolated claim on a hard-to-parse page is a weak candidate; a fact stated plainly, documented with primary sources, and echoed across credible outlets is a strong one. That is the logic behind the profile work described on our About page — though no effort can guarantee that any platform will cite any particular source, and this site makes no such promise.

What this section covers

Questions this section answers

Readers will be able to answer: Why does one platform cite sources while another gives an answer with no links? Does being cited mean the AI system agrees with the source? How should a business read a report about its own citation share? And what is the difference between being named in an answer and being used as its evidence? The retrieval mechanics behind these questions are covered in How AI Retrieves Information, and the question of what a business can do about its own visibility is taken up in Building AI Visibility and Authority.

How it connects to the other sections

Citations are the visible end of a chain that starts with how models generate answers and passes through retrieval. Entities, Knowledge Graphs and Semantic SEO explains how a business’s identity is represented in the systems that decide which sources are relevant. AI Visibility Measurement and Research explains how citation tracking is actually done, and what its limits are. The full overview is on the Knowledge Center page.

What comes next

This introduction frames the subject; later articles will document each platform’s citation behavior in detail and explain citation measurement methods step by step. If you notice an error, or can point to a source that improves an explanation, please use our contact and corrections form.

Latest in this pillar