Module 02 · Lesson 3 of 4
AI Citations and Source Eligibility
An AI citation is a reference or link displayed alongside a generated answer. Different surfaces display them differently: inline markers, a link panel, a sources list. The display varies; the meaning is stable. Somebody's material got used, and that somebody is now visible to the reader at the exact moment of the question.
Two different problems
Eligibility asks: could your content be used at all? That is a hygiene problem. Google is explicit that no special schema or extra technical work makes you eligible; ordinary indexed, crawlable, well-built pages qualify. Being blocked, thin, or duplicated is how sites fail eligibility without ever knowing it.
Citation asks: of all eligible sources, why did the answer use that one? That is a competition problem, and it is decided by relevance and quality of the passage, the credibility of the source, and corroboration across the web. No markup trick solves a competition problem. This distinction kills half the bad advice in the AI SEO space on its own, including the persistent myth that adding schema causes citations. It does not. Google says plainly that AI features require no special schema.
What tends to get cited
Observed across many queries, the sources that recur in generated answers share a few traits. They are primary for the claim being made (official docs for platform facts, the study itself for research claims). They are specific rather than generic. They are structured so their key passages stand alone. And they are corroborated: other credible sources reference them. Treat that list as observed industry consensus plus reasoned inference, not a leaked specification. The way to know what gets cited in your niche is to look, which is exactly what the workbench does.
| Source type | Typically cited for | Example |
|---|---|---|
| Official documentation | Platform facts and rules | Google Search Central on AI features |
| Primary research | Statistics and findings | The original study, not a blog retelling it |
| Specialist publications | Explanations and analysis | Trade press in the niche |
| Comparison or data pages | Options and prices | Structured comparison pages |
| Communities and forums | First-hand experience | Detailed practitioner discussions |
- Trigger any AI answer for a question in your niche (an AI Overview, an assistant response, anything with visible sources).
- List every cited domain in your log.
- Label each with a type from the table above, or invent a label if none fits.
- Ask one question per domain: what does this source have that mine does not? Primary status, structure, specificity, brand, corroboration. Write the answer down.
- Take three queries from your Workbench 2.1 log that produced AI answers.
- For each, record every cited domain in a sheet with columns: query, domain, type label, your-domain-or-not.
- Count the types. Note which type dominates your niche's answers.
- Star every recurring domain. Those are your citation competitors, and your Module 6 prospect list.
Does adding structured data make an AI system cite you?
Your site is crawlable, indexed and well built, but never cited. Which problem do you have?
Why does primary status matter so much in citations?
Sources used in this lesson
Google Search Central: AI features and your website
Google Search Central: structured data