What Is a Cited Source?

A cited source is a web page that an AI engine references — usually with a link — when constructing a generated answer. It is the page that survived the full pipeline: crawled by an AI agent, indexed, retrieved for the query, ranked highly by the reranker, and trusted enough to build a claim from. Being a cited source is the concrete, countable outcome that GEO optimizes for.

What attributes do cited pages share?

Cited pages consistently exhibit three properties. They are retrievable — indexed and crawlable by agents like OAI-SearchBot, PerplexityBot, and Googlebot. They are extractable — written in answer-first passages that make sense quoted in isolation. And they are evidenced — dense with specific numbers, dates, and named sources the engine can lift verbatim.

Why do engines prefer some sources over others?

• Specificity: a page with an exact statistic beats a vague one, because the model can attribute a precise claim.
• Structure: clear H2 questions and tables give the reranker clean chunks to score.
• Authority signals: named authors, an about page, and sameAs links reduce the risk of citing a low-trust source.
• Freshness: recently updated pages win time-sensitive queries.

The GEO paper (Aggarwal et al., KDD 2024) found that adding quotations, statistics, and citations lifted a page's generative visibility 30-40%, meaning citability is engineered, not accidental.

How is a cited source different from a linkless mention?

A cited source carries an attribution the engine exposes, typically a link. A linkless mention names your brand in the answer text without pointing at a page. Both build authority, but only the cited source proves your specific content entered the pipeline.

Example

A logistics analyst publishes "average last-mile delivery cost per parcel, 2026" as a stand-alone data page with a table and method note. Because it is the cleanest primary source, Perplexity and ChatGPT both cite it — the page becomes the category's default attribution for that number.