AI Overviews

An AI Overview is a featured snippet that writes its own copy. Google’s model reads the candidates, composes a summary, and cites a handful of links. For the user it is often genuinely faster. For the sites below it, the math changed: the overview satisfies the click that the snippet only teased.

Where the clicks actually go

The panic version — “AI Overviews kill SEO” — does not survive contact with query-level data. The effect splits by intent:

• Answer-shaped queries (“what is”, “how does X work”, comparisons) lose heavily. If the overview answers it, the click was optional all along — this is the zero-click curve steepening, not a new species.
• Do-and-buy queries keep their clicks. Nobody replaces a purchase with a summary.
• Cited sources keep a trickle that compounds. Overview links send less traffic than position 1 did, but the citation also propagates into every downstream AI that reads Google, and into user trust. Being the source of record outlasts any single SERP layout.

The citation pattern

Pages that earn overview citations are not lucky. They answer the question in the first screen, in plain words, with the specifics attached — a date, a number, a source, a first-person test. Original data is disproportionately cited because the model needs somewhere to ground each claim. Fifth-hand restatement of the same paragraph has nothing to offer a synthesizer.

This is also why GEO scoring weights facts and structure so heavily: those layers are what extraction and citation run on. The practical checklist is short — say the answer, prove it, structure it, sign it — and it is the same checklist that survived every previous SERP shakeout.

What not to do

Do not delete your informational content because overviews answer it. The citations have to come from somewhere, and abandoned topics get summarized from worse sources. Do measure per-query: your GSC data will show exactly which pages lost clicks to a feature and which did not. And do not rebuild your site around an unindexed chatbot wall — the overview’s sources are still web pages, still ranked by the machinery you already optimize for.

FAQ

What are AI Overviews?

AI Overviews are Google's generated answer summaries shown above traditional results for qualifying queries, with links to cited sources. They extend the featured-snippet pattern from extractions to synthesis: the overview is composed by a model from multiple sources, not quoted from one.

Do AI Overviews kill organic traffic?

They compress it selectively. Informational, definitional and how-to queries lose the most clicks — the overview answers them. Queries with transactional intent, local intent, or a need for fresh/specific data keep their clicks, and links inside overviews still send traffic. The damage is real for answer-shaped content and minor for the rest.

How do you get cited in an AI Overview?

The cited pages share the pattern: they answer the question directly near the top (BLUF), state facts with dates and sources, use clean structure (headings, lists, tables), and demonstrate first-hand experience. Being the origin of a fact or a number — not the fifth site repeating it — is the strongest citation signal.

Should you optimize differently for AI Overviews vs classic SEO?

Less differently than the discourse suggests. The same fundamentals — clarity, structure, verifiable claims, topical depth — drive both. The additions are citation-shaped: original data, explicit answers, schema where it aids extraction, and an llms.txt layer for the non-Google engines.