How to appear in Google AI Overviews

Google AI Overviews pull from the same index and largely the same ranking signals as classic search, then layer on passage-level extraction — so you need to rank reasonably well first, and then be written in a way that a specific passage can be lifted cleanly into a summary. There's no separate submission process; it's an extension of good SEO plus answer-first structure.

Is ranking in AI Overviews the same as ranking in classic search?

Mostly, yes, with an added filter on top. AI Overviews draw their source pages from Google's normal web index, so the usual fundamentals still apply: crawlability, relevance, backlinks, page experience, and topical authority. A page that can't rank on page one for a query is very unlikely to be pulled into the Overview for that same query.

The added layer is passage-level extraction. Google isn't just deciding whether your page is relevant overall — it's identifying a specific passage or sentence that answers the question well enough to quote or paraphrase in a compressed summary. You can rank respectably and still be skipped for the Overview if no single passage on your page states the answer cleanly enough to lift.

So the practical approach is two jobs stacked: do the SEO work to earn a ranking position, then structure the content so a self-contained passage exists for Google to pull from.

What is passage-level quotability, and how do you write for it?

It means a single sentence or short block of text answers the implicit question completely on its own, without needing the rest of the page for context. Google's summarisation systems favour passages that are already close to the ideal answer, since less rewriting means less risk of misrepresenting the source.

To write quotable passages:

• Put the direct answer in the first sentence of the relevant section, then explain or qualify afterward.
• Keep the answer sentence factual and specific — a number, a name, a defined process — rather than a vague generalisation.
• Avoid burying the answer inside a long compound sentence with multiple clauses; short and direct extracts more cleanly.
• Repeat the core fact near the top of the section even if it's expanded on later, since extraction tools weight early content in a section more heavily.

Does Google-Extended control whether you appear in AI Overviews?

Not entirely — this is a common point of confusion. Google-Extended is the user agent that controls whether your content can be used to train Gemini models; it is a separate control from whether Googlebot can crawl and index your pages for search, which is what actually feeds AI Overviews.

• Blocking Google-Extended does not automatically remove you from AI Overviews — that's governed by your normal robots.txt rules for Googlebot and your indexing status.
• Blocking Googlebot, or having noindex tags left over from staging, will remove you from both classic results and Overviews.
• If you deliberately want to opt out of Gemini training but stay eligible for search and Overviews, disallow Google-Extended specifically and leave Googlebot access untouched.
• Check both settings separately; they get conflated often enough that sites end up blocking the wrong one.

How much does schema markup actually help?

Schema doesn't directly cause an AI Overview to appear, but it makes your content far easier for Google's systems to parse unambiguously, which reduces the chance of misinterpretation and supports eligibility for rich extraction generally. FAQPage and Article schema in particular give a structured, low-ambiguity version of the same answer-first content you should already be writing.

Schema should describe what's already visible on the page accurately — mismatched schema and visible content is a trust red flag, not a shortcut.