Optimising for Google AI Overviews

Google AI Overviews are AI-generated summaries that appear at the top of the search results, above the traditional links, answering the query directly and citing a handful of sources. For many searches they are the first thing a user reads, and often the only thing.

AI summaries now appear on roughly half of searches, a share McKinsey expects to pass 75% by 2028.

That makes the Overview the most valuable position on the page. Being cited in it puts your brand inside the answer itself; being left out means competing for the shrinking attention that reaches the links below.

It draws on its main search index, then uses a query fan-out to break the question into sub-queries and assemble an answer from the strongest passages it finds. So classic search visibility feeds AI Overviews, being crawlable and indexed is the price of entry, but it does not decide the outcome. The mechanism is explained in query fan-out.

The surprising part is how loosely inclusion tracks ranking.

In one analysis of Google’s AI answers, a large majority of cited pages did not rank in the organic top 10.

That figure is from a 2026 analysis, so treat the exact share as directional, but the implication is robust and repeated across studies: Google is choosing the clearest, most liftable, most trustworthy passage for each sub-question, not simply the highest-ranked page. Extractability can beat rank.

Because the Overview answers the question, so fewer people scroll. The effect on traditional traffic is significant and well documented.

When an AI Overview appears, the top organic result loses about 58% of its clicks (Ahrefs), and users click a traditional result just 8% of the time versus 15% without a summary (Pew).

This is the uncomfortable trade at the heart of AI Overviews: the visibility is enormous, but the click is smaller. It is the reason to compete for the citation rather than only the rank, and the reason to measure success by presence in the answer, not just by traffic.

By being eligible, liftable, comprehensive and trusted, in that order. The moves build on each other.

Keep a strong, crawlable foundation. Be indexed, fast and technically sound. You cannot be cited in an Overview if Google cannot crawl and understand your page.

Structure for a clean lift. Lead with a direct answer, use question-shaped headings and short self-contained sections, so Google can extract a passage cleanly.

Cover the sub-questions. Because fan-out fragments a query, breadth wins. A topic covered from every angle is eligible for more of the sub-answers than a single narrow page.

Add evidence and freshness. Statistics, cited sources and a recent update date all raise the odds of being the passage Google chooses.

Earn off-site trust. Mentions and coverage in sources Google trusts make it more comfortable quoting you.

Yes. The click is smaller, but the alternative is worse. If the Overview answers the question and names three competitors instead of you, you lose the visit and the consideration. Being cited keeps your brand in front of the buyer at the decisive moment, shapes the shortlist, and captures the smaller but higher-intent click that remains. Absence from the answer is the real cost.

Not cleanly, and it is rarely wise. Blunt controls such as the nosnippet directive can limit how your content is used, but they tend to reduce your visibility across normal search too, so they trade one problem for a larger one. For almost all brands the better path is to compete for the citation rather than hide from it. Confirm any technical directive against Google’s current documentation. [verify snippet controls against current Google docs]

Ranking gets you eligible; extractability gets you cited. Keep a strong search foundation, then structure a clean answer, cover the whole topic, and earn the trust that makes Google comfortable quoting you in the summary above the links.