Citations and Sources
The source mix is wider than your site#
On a typical buyer prompt the cited material spreads across several types, and owned content is only a slice of it:
| Source type | Why models lean on it |
|---|---|
| Community threads | Unfiltered peer opinion, heavy on real trade-offs |
| Review sites | Structured comparison, ratings, recency |
| Comparison pages | Directly answer "X vs Y" questions |
| Official docs | Authoritative on capability and specifics |
| Owned content | Positioning, features, pricing |
The practical consequence: a perfectly written homepage does not carry an answer on its own. Presence has to exist in the places the model already trusts for your category.
What the platform records#
For every tracked answer, AI Search Console stores the domains and URLs cited, the source type, how often each source appears across your prompt set, and which answers each source influenced.
That last one turns a list into a priority order. A source cited in one answer is trivia. A source cited in most of your category prompts is leverage.
Reading a citation list#
Three patterns come up constantly:
A source cites your competitor and not you. The most fixable case. The page exists, the model trusts it, you are simply not in it.
A source cites you inaccurately. Worse than absence, because the model repeats the error to every buyer. Correct it at the source.
A high-frequency source you have never touched. Usually a community thread or a comparison page you did not know was carrying your category.
Next#
- Find Citation Gaps — the workflow built on this data
- Share of Voice