How to appear in Google AI Overviews

AI Overviews are the one AI surface where Google has written down the rules, at least partly. Its Search Central documentation on AI features says there are no extra technical requirements. A page needs to be indexed and eligible to show in Search with a snippet. The usual SEO fundamentals apply, and there's no special markup, file, or opt-in that gets you included.

That's useful, because it rules out a lot of snake oil. It's also frustrating, because "do good SEO" isn't a plan. Here's how we turn it into one.

Step 1: confirm the page is eligible at all

Start with the boring checks, because they fail more often than you'd think.

Indexed. Use the URL Inspection tool in Search Console on the pages you want linked. If a page isn't in the index, nothing else matters.
Snippet controls. Google's snippet directives also apply to AI features. A nosnippet robots meta tag, a tight max-snippet value, or data-nosnippet on the key section limits what can be shown. Teams sometimes add these for legal or pricing reasons and forget about them.
Rendering. Googlebot renders JavaScript, but rendering problems still happen. Check the rendered HTML in URL Inspection and make sure the answer you want quoted is in it.

One thing that doesn't matter here: Google-Extended. It's a robots.txt token that controls use of your content for Gemini apps and Vertex AI. Google says it doesn't affect inclusion in Search, and AI Overviews are part of Search.

Step 2: find which of your queries trigger an Overview

AI Overviews don't show for every search. Google shows them when its systems judge they add something beyond the normal results, so many navigational and simple queries never get one. Before you rewrite anything, find out which queries in your category produce an Overview and which don't.

Work from your existing keyword list and the questions buyers ask in sales calls. Run them in a clean browser session, in the country you sell in, and note three things for each: whether an Overview appeared, which pages it linked, and which brands it named in the text. Expect the result to vary. The same query can show an Overview one day and not the next, and the linked sources rotate.

Step 3: understand query fan-out

Google's documentation says AI Overviews and AI Mode may use a technique it calls query fan-out: issuing several related searches across subtopics and then building the response from what comes back. This is why an Overview often links pages that don't rank on page one for the original query. They ranked for one of the narrower searches behind it.

For you, it means the target isn't only the head query. It's the set of subtopics the head query breaks into. For "best accounting software for freelancers", those might cover multi-currency invoicing, tax handling in a given country, pricing, and how a specific product compares with the usual alternatives. A page that answers one of those subtopics clearly has a shot at a link even if a much bigger site owns the head term.

Step 4: write for the subtopic, then for the snippet

Look at the pages that got linked and you'll usually see the same habits. They answer a narrow question directly, near the top of a section, in a sentence or two that makes sense out of context. They state specifics. They're structured with descriptive headings, so a passage can be pulled without dragging the rest of the page along.

So for each subtopic you found in step 3:

Decide which page on your site should own it. One subtopic, one page, so you're not splitting the signal.
Add a heading that matches the question and a short direct answer beneath it.
Back the answer with detail a reader can act on: steps, constraints, a table where comparison is the point.
Keep it honest and current. Google's guidance keeps coming back to helpful, people-first content, and a stale fact is an easy reason to link someone else.

Step 5: treat structured data as hygiene

Google supports structured data in Search, and its AI features guidance says markup should match the visible content on the page. It doesn't say structured data gets you into AI Overviews, and we don't sell it that way. Add the schema types that fit your templates because they help Google understand the page and power rich results. Don't expect them to override weak content.

Step 6: measure what Search Console can't show you

Search Console is the official source for clicks and impressions, and Google says appearances in AI features are counted inside your normal Web search performance data. That's handy for totals. What it won't show you is the Overview itself: the text Google wrote, the brands it named, and which competitors were linked next to you.

That's the gap we fill with Promptwatch. We track the queries from step 2 as prompts, and Promptwatch records the Google AI Overview and AI Mode answers alongside ChatGPT, Gemini, Perplexity, and the rest. Citation analytics then shows which URLs each Overview linked, page by page, and citation trends show whether your pages are gaining or losing those links over time. Query fan-outs are visible on tracked prompts too, which makes step 3 a lot less guesswork.

The two data sources work best side by side. Search Console tells you whether clicks moved. The prompt data tells you why: whether you were added to an Overview, dropped from it, or still named in the text but no longer linked.

A realistic cadence

Pick twenty or thirty queries that trigger Overviews and matter commercially. Fix eligibility first, then work through the subtopics one page at a time. Recheck the tracked prompts after each batch of changes has been recrawled, not the next morning. AI Overviews move, but they reward the same patient, specific work that good SEO always has.