Position one is not the top anymore.
The citation above it is.

Standard SEO competes for a position in a ranked list. An AI Overview is written from a retrieved set of sources, which is much shorter and selected on different grounds. Ranking well makes a page a candidate for that set. Whether it gets quoted depends on how cleanly one passage can be lifted and stay accurate, how unambiguously your entities are defined, and how visible the trust signals behind the claim are.

The practical difference is where the leverage sits. In classic organic work, links do a lot of it. Here, structure, markup, authorship and entity clarity do more, which is why a site with a lighter link profile can be cited above the page outranking it.

Google has not published a source-selection algorithm, so anyone stating the rules with certainty is guessing. What is observable across live queries is a consistent pattern: the cited pages tend to answer the specific question directly and early, make claims concrete enough to attribute, carry recognizable authorship, describe their entities the same way everywhere, and hold together when a passage is pulled out of context.

Those are the levers, and they are the ones we work on. Where something is inference rather than documentation, the audit says so rather than presenting it as a rule.

No. A featured snippet quotes a single page close to verbatim. An AI Overview is generated text written in Google's own words from several retrieved sources, with links out to them, which means more than one site can be cited on the same query.

Winning a snippet is about being the clearest single answer. Being cited in an overview is about being one of several usable sources. The structural groundwork overlaps enough that we target both, but optimizing for one does not hand you the other.

An entity is a thing Google can identify and hold facts about: your company, your founder, a service, a location. Generated answers are assembled around entities rather than keyword strings, so a brand that exists only as a phrase on a page is harder to reference than one that resolves to a known thing.

In practice it means naming and describing those things consistently across the site, marking up the relationships so nothing is left to inference, linking out to the external profiles Google already trusts, and closing the gaps that keep a Knowledge Graph entry from forming. Semantic search rewards the same clarity, which is why this work tends to lift ordinary organic performance alongside citation. It is also the same entity groundwork Answer Engine Optimization depends on for other assistants.

Individual pages sometimes get pulled in within a few weeks of being restructured, because overviews refresh faster than rankings and do not need a position change to start citing you. Consistent presence across a set of queries usually takes a few months, since entity and E-E-A-T signals accumulate rather than switch on.

We log the query-level starting position before anything changes, so progress is measured against a recorded baseline rather than described after the fact.

By checking the queries and recording what comes back. For each priority query we log whether an overview appears, which sources are cited, and where you sit among them, then recheck on a schedule. No tool reports citation presence reliably yet, so this part is done by hand and dated.

Search Console impressions and clicks sit next to it in the same report, because a page can gain a citation and hold its organic traffic at the same time, and both numbers belong in the same view.

This page is about Google's own AI Overview feature. Coverage across ChatGPT, Gemini, Perplexity and other assistants is our GEO service, a separate audit with its own retrieval behavior and its own crawler access questions to account for.

The two share a lot of groundwork and most clients scope them together, so they usually run side by side. They are still two engagements, and we do not price one as though it delivers the other.