Does schema markup help AI Overviews?

01Does schema markup help AI Overviews? The short answer

Does schema markup help AI Overviews? Not in the way the question usually means. Google's own documentation says there is no special markup that gets a page into AI Overviews or AI Mode, and it says so twice, on two different pages. What markup does is help Google understand what a page is about, which is useful everywhere, including the systems that pick supporting links for an AI answer.

Schema markup is structured data written in the schema.org vocabulary, usually as a JSON-LD block, that labels the things on a page: the article, its author, the product, the rating. An AI Overview is the generated summary Google shows above some search results, with links to the pages that support it. The two are related, but not by a switch you can flip.

This guide separates what Google documents from what is claimed in the AEO and GEO advice market, and ends with a short list of what to do. If you are new to the topic, our overview of markup that machines read explains structured data from the ground up, and the AI answers section of this handbook collects every guide on how answer engines treat markup.

02What does Google say about structured data and AI features?

The clearest statement is in Google's page on AI features and your website. It says there are no additional requirements to appear in AI Overviews or AI Mode, and no other special optimizations necessary. To be eligible as a supporting link, a page must be indexed and eligible to be shown in Google Search with a snippet. The same page adds that you do not need to create new machine readable files, AI text files or markup to appear in these features.

In May 2026 Google published a longer guide to optimizing for generative AI features on Google Search. Its mythbusting section lists "overfocusing on structured data" among the things you can ignore. Structured data, it says, is not required for generative AI search and there is no special schema.org markup to add. The same sentence goes on to recommend that you keep using structured data as part of your overall SEO, because it helps with being eligible for rich results.

Read together, the two pages give a consistent position as of November 2026:

• Markup is not an entry ticket. A page with no structured data at all can be cited in an AI Overview if it is indexed, snippet eligible and the best source for the question.
• Markup is not a booster. Google does not describe any ranking or citation bonus for AI features that comes from having JSON-LD.
• Markup is still recommended. It remains part of normal SEO, it powers rich results, and Google uses it to understand pages and the wider web.
• Markup must be honest. The AI features page lists "making sure your structured data matches the visible text on the page" among the SEO fundamentals that still apply.

That last point connects to Google's introduction to structured data, which says Google uses the structured data it finds to understand the content of a page and to gather information about the web and the world, such as the people, books or companies in the markup. Understanding is the honest scope of the benefit. AI Overviews are built on the same index and ranking systems, so anything that helps Google understand a page can help indirectly. Nothing in the documentation goes further than that.

The optimization guide adds one practical detail that is easy to miss. Beyond the normal technical requirements, a site must be included in Search generative AI features in Search Console to be eligible for display in those features. Google also points to a Generative AI performance report in Search Console as the way to measure how content performs there. Both are settings and reports, not markup.

03What do controlled tests show?

Less than the marketing around them suggests. Google has not published a controlled experiment on schema markup and AI Overview citations. The case studies Google does publish on its structured data introduction page are about rich results and click-through rates, from years before AI Overviews existed, so they cannot be read as evidence about AI answers.

Independent SEO experiments on this question circulate widely, and their results point in different directions. Most are small, use a handful of pages, and change the markup at the same time as other things. This handbook does not cite figures from them, because none we could trace met the bar of a published method that someone else could repeat.

If you want to run a fair test on your own site, the design matters more than the tool. A controlled test of markup and AI citations needs at least these conditions:

1. Pairs of comparable pages on the same site, with similar topics, age, links and traffic, so that markup is the only meaningful difference.
2. A change to the markup alone, with the visible text, title and internal links left untouched for the length of the test.
3. A fixed query set checked on a schedule, because AI Overviews often do not trigger and the links they show vary from one run to the next.
4. Enough time for recrawling, which Google's AI features page says can take from several days to several months.
5. A written record of every observation with its date, so a null result is reported as plainly as a positive one.

Even a well run test on one site tells you about that site. The honest summary today is that no public, repeatable evidence shows schema markup on its own causing AI Overview citations, and Google's documentation says not to expect it.

04Where does markup clearly help?

Markup earns its place where it describes things that are hard to infer from prose, or where Google ties a search feature to it. Four areas stand out.

Entities: who and what the page is about

Organization, Person and sameAs markup tell Google which real organization publishes a site and which real person wrote an article. Google's structured data introduction says it can make general use of the sameAs property and other schema.org structured data beyond the documented rich results. The guide to how ChatGPT and Perplexity read your pages covers what other engines are documented to do with the same markup.

Products and local businesses

Google's AI optimization guide says generative AI responses can include product listings, product information and local business information, and it points site owners to Merchant Center feeds and Business Profiles to keep that data visible. On-page Product markup complements a feed; it does not replace it. For a store, the feed and the page should state the same price and availability.

Reviews and ratings

Review and AggregateRating markup still powers review snippets for a defined list of types. It also gives any system reading the page an exact, machine-readable rating instead of a number buried in a sentence. The rating must be real and visible: Google's structured data guidelines warn that reviews or ratings not by actual users may result in a manual action.

Rich results that still exist

Rich results such as product snippets, review snippets, breadcrumbs and organization logos are the documented payoff of structured data in Google Search. Google's general structured data guidelines still govern them: mark up only content that is visible, relevant and complete. To see what your markup actually outputs, follow the steps in testing structured data the right way.

05Where markup makes no visible difference

Some of the most repeated AI search tips do nothing for Google, by Google's own account. The mythbusting section of the AI optimization guide names several:

• llms.txt and other AI text files. Google Search does not use them. Keeping one for other services is fine, and it neither helps nor harms Google visibility.
• Special AI markup. There is no schema.org type that marks a passage as an answer for AI Overviews.
• Chunking content into tiny pieces. Google says its systems can understand multiple topics on a page and show the relevant part.
• Rewriting just for AI. AI systems understand synonyms, so stuffing every phrasing of a question adds nothing.

For llms.txt in general, the explainer on robots.txt and llms.txt covers what the file is and who reads it.

Google is also the only major answer engine that documents its position on markup in this much detail. ChatGPT and Perplexity publish documentation about their crawlers, not about how a fetched page is parsed. Claims that one of them reads a particular schema type more closely than another have no published basis, which is why the crawler guide linked above sticks to what each product documents.

06What should you do now?

Treat structured data as accurate labelling, not as an AI tactic. The work that pays is the same work that has always paid, done carefully.

1. Keep pages indexable and snippet eligible. A nosnippet rule or a blocked crawler removes a page from AI features entirely.
2. Mark up what the page really is: Article with a real author and honest dates, Product with the visible price, Organization once for the site.
3. Match every marked-up value to visible text, and test after every theme or plugin update.
4. Write passages that can be quoted on their own. The checklist for quotable pages shows how direct answers, tables and question headings help.
5. Drop tactics that Google says it ignores, and spend the time on content that adds something the rest of the web does not already say.

Publishers who want a newsroom view of AI Overviews can read how publishers approach Google AI Overviews, which covers editorial strategy rather than markup.