How do you optimise content for Google AI Overviews?

Google's guidance on its AI features is short:

• The page must be indexed and eligible to appear in Search with a snippet.
• There are no extra technical requirements and no special schema for AI Overviews or AI Mode.
• The normal snippet controls apply. nosnippet, data-nosnippet and max-snippet limit what can be shown, and noindex removes the page entirely.
• Google-Extended, the robots.txt token for Gemini training, does not control whether a page appears in Search or its AI features.

So the first check is mechanical: in Search Console's URL Inspection tool, is the page indexed? Does it carry a snippet restriction someone added and forgot?

AI Overviews are generated from pages in Google's index that are relevant to the query, and they show supporting links to those pages. AI Mode goes further: it runs several related searches for one question (Google calls this query fan-out) and combines what it finds. That means:

• Sub-questions are opportunities. A page that clearly answers "how long does it take" or "what documents do I need" can be cited inside an answer to a broader question.
• Passages compete, not whole pages. A clean two-sentence answer inside a long guide can be the bit that gets used.
• Trust carries over. Pages from sites Google already ranks well on a topic are more likely to appear. AI Overviews are not a shortcut around authority.

1. Search the query yourself. Note whether an AI Overview appears, which sources it cites and what sub-questions it covers.
2. Write the opening answer. Two to four sentences that answer the main question directly, with the key specifics.
3. List the follow-up questions. Use the AI Overview, People Also Ask and your own sales calls. Give each one a question-led H2 or H3 with a direct answer in its first sentence.
4. Format for the content type. Numbered steps for processes, tables for comparisons, short definition sentences for terms.
5. Add evidence. Specific ranges, named regulations or standards, examples from your own work, and a named author or team.
6. Check structured data. Article, Organization, Product or LocalBusiness schema that matches the visible content. Validate it with the Rich Results Test or Schema Markup Validator.
7. Link internally from related pages so Google reaches and understands the page in context.
8. Request re-indexing in Search Console after significant changes, then re-check the query in a few weeks.

Since 2023 Google shows FAQ rich results only for a limited set of well-known government and health sites, and it has dropped HowTo rich results. Adding FAQPage or HowTo markup will not bring back those visual features for most sites. Visible FAQ sections are still useful, because they answer follow-up questions in a quotable format. Mark them up only if the markup matches the page, and do not expect a rich result.

• Burying the answer under a long introduction about why the topic matters.
• Restricting snippets with a max-snippet value or nosnippet added for other reasons, which can limit how the page is used.
• Schema that does not match the page, which Google ignores at best.
• Writing for the engine, not the reader. Stilted, repetitive Q&A copy hurts conversions from the people who do click.
• Expecting traffic like a top organic result. AI Overviews answer the question on the page. Some users still click, often the more qualified ones.

Here is a hypothetical. Say you run a company formation consultancy in Dubai, and "how to set up a company in a Dubai free zone" shows an AI Overview citing two government portals and a competitor.

Your existing page opens with three paragraphs about Dubai's business environment. You rewrite it: the first paragraph lists the main steps in one sentence each; separate sections answer the sub-questions the AI Overview covers (choosing a free zone, licence types, visa allocation, timelines, typical cost components); a table compares three free zones on the factors buyers ask about; and the page names the consultant who wrote it. You check the page is indexed with no snippet restrictions, request re-indexing, and recheck the query monthly.

We start with the queries that matter commercially and check each one in Google for AI Overview and AI Mode presence, cited sources and the sub-questions covered. Then we audit your matching pages for eligibility (indexing, snippet controls, rendering) and for extraction (where the answer sits, which sub-questions are missing).

The fixes are page rewrites, new sections or new pages for uncovered sub-questions, and schema that matches the content. If your CMS cannot output clean HTML and schema, or pages render content only through scripts, a rebuild through Web Design is often faster than workarounds. Results go into the same AI visibility report as ChatGPT and Perplexity, so you can see all engines in one place.

Get my AI visibility check →