Is There a Difference Between Ranking in Google and Being Cited by an AI?

What does Google require for a page to be cited in its AI features?

Nothing beyond what it already requires to appear in search. Google's documentation on AI features is unusually direct: “To be eligible to be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet.”

Both conditions are ordinary. Indexing is the baseline requirement for appearing in Google at all. Snippet eligibility means the publisher has not used the controls that suppress text extracts. A page meeting the normal conditions for search is, by that fact, eligible for the AI surfaces.

The word doing the work is again eligible. Meeting the conditions makes a page available for selection; it does not cause selection, and Google does not publish how the choice among eligible pages is made.

This matters commercially because a whole category of services has grown up around the premise that AI citation is a separate discipline requiring separate work. On Google's own account of its own products, the entry requirement is the search pipeline that already existed.

It is worth noting what the requirement does not mention. There is no stipulation about content format, no markup obligation, no minimum length, no publication tier and no registration of any kind. The gate is the index, and the index is open to anyone with a page Google can fetch.

Is there such a thing as AI optimization?

For Google's features, the company denies it in terms: “There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.”

That sentence is worth reading twice, because it is aimed squarely at a fast-growing market. Services sold under names like answer engine optimisation or generative engine optimisation propose that AI answers respond to techniques distinct from search. Google's documentation says that for its own AI surfaces, the foundational practices are the same ones that apply to Search generally.

Nothing here says the underlying selection is identical to classic ranking, and it plainly is not — the presentation differs and so does the set of links shown. The claim is narrower and more useful: there is no separate optimisation lever, and no work a publisher can do that targets the AI surface specifically.

So a proposal to make a company “visible in AI answers” through special technical treatment is, for Google's products, selling work its operator says is unnecessary. The honest version of that service is ordinary search work, which is a legitimate thing to buy under its own name.

The market pressure behind the claim is easy to see. A new surface appears, buyers assume new rules, and a service category forms to sell mastery of them before anyone can check. Google publishing a denial does not dissolve that market, because the denial is less exciting than the promise and considerably less profitable to repeat.

What controls do publishers actually have?

The ones they already had. Google points publishers to nosnippet, data-nosnippet, max-snippet and noindex to limit how content is displayed, and separately to Google-Extended for limiting “AI training and grounding in some of Google's other systems”.

The design has a consequence publishers sometimes discover late. Because AI eligibility is tied to snippet eligibility, suppressing appearance in AI Overviews by removing snippets also degrades the ordinary search listing, which is the same trade-off that has always applied to snippet controls.

Google-Extended is the separate lever, and it addresses training and grounding rather than whether a page can be cited as a supporting link in a search result. Conflating the two produces a common error: a publisher blocks Google-Extended believing it has removed itself from AI answers in Search, which is not what that control does.

For a placement buyer the implication is modest but real. Whether coverage can be surfaced in these features depends on the publisher's configuration, not on anything the buyer arranges, and it is a property of the outlet worth asking about rather than assuming.

How does query fan-out change which pages appear?

Google describes its AI features as using a “query fan-out” approach, issuing multiple related searches behind a single question rather than one, and says the result is that these features “display a wider and more diverse set of helpful links associated with the response than with a classic web search”.

If that holds, the practical effect is not that different pages qualify but that more of them surface. A page ranking modestly for the exact phrase a user typed may still be retrieved by one of the related queries generated behind the scenes, and appear as a supporting link it would not have earned in a classic result.

This is the strongest available argument for coverage that answers specific, narrow questions. Fan-out rewards documents that resolve a sub-question precisely, because the sub-questions are what the system is searching for.

It is worth marking the limit of this reasoning. Google describes the mechanism and the diversity of links; it does not publish how candidates are ranked or chosen within it. The inference that narrow, specific coverage benefits is plausible and follows from the described design, and it is not a published finding.

What about AI systems Google does not operate?

Nothing above transfers, and the difference is the most important caveat in this report.

Google's willingness to document eligibility for its own features is unusual. Other assistants and answer engines do not publish comparable accounts of how sources are selected for citation, which is why the question “how do I get cited” has a partial answer for Google's surfaces and no documented answer elsewhere, as examined separately in this series.

What can be said generally is structural. Being crawlable is a precondition everywhere; whether a crawler visited, retained, retrieved and attributed a page involves steps no operator publishes. A document that cannot be fetched cannot be cited, and a document that can be fetched has no guarantee of anything.

The practical consequence is that claims about AI visibility should be checked against which system is meant. A vendor citing Google's documentation to support a promise about a different company's assistant is transferring evidence across a boundary the evidence does not cross.

This asymmetry also explains why so much advice in this area is transferred without attribution. Google's documentation is the only detailed public account available, so it becomes the evidence base for claims about every other system, none of which has agreed to be described by it.

What this means for anyone buying placement

Stop paying a premium for AI optimisation as a distinct service, at least where Google's features are the target. Its operator states there are no additional requirements and no special optimisations necessary, in documentation that is free to read.

Ask what is actually being sold. Ordinary technical search work — indexable pages, accurate markup, content that answers real questions — is worth buying and is the same work that makes a page eligible here. A premium charged for the AI framing buys the label, not additional mechanism.

For coverage specifically, the useful question is whether the publication allows snippets and is indexed, because those are the conditions that decide eligibility. A placement on an outlet that suppresses snippets is a placement that cannot be surfaced this way, whatever was promised.

Media Morpheus sells placement and search visibility work. What that can produce is an indexed, snippet-eligible page answering a specific question, which is the documented eligibility condition for Google's AI surfaces and the whole of what anyone can arrange. Selection among eligible pages is Google's, and no vendor has access to it.

Methodology and limitations

This report relies on Google's published documentation covering AI features in Google Search, quoted for its statements on eligibility, optimization, publisher controls and query fan-out. Four limits apply. First, the documentation describes eligibility rather than selection: Google does not publish how it chooses among eligible pages, so nothing here explains why a particular page is cited. Second, findings apply only to Google's own AI surfaces; other assistants and answer engines publish no comparable documentation and no statement here should be read as describing them. Third, the discussion of query fan-out and narrow questions is an inference from Google's described design, not a published finding, and is marked as such in the text. Fourth, platform documentation changes without notice and this describes the position at the date of publication.

Conclusion

For Google's own AI features the answer is close to no: a page becomes eligible for citation by being indexed and snippet-eligible, the conditions that already governed search, and Google states there are no additional requirements or special optimizations. That undercuts the premise of a service category currently being sold at a premium. What remains true is that eligibility is not selection, that Google publishes the first and not the second, and that none of this documentation describes AI systems Google does not run. The work that makes a page eligible is ordinary, well understood, and worth doing under its own name rather than a newer and more expensive one.

Frequently Asked Questions

Sources

1. official Google Search Central. Documentation on AI features in Google Search, covering eligibility for AI Overviews and AI Mode, publisher controls, and query fan-out.
2. official Google Search Central. Documentation on Google-Extended, for the distinction between search citation and AI training or grounding.