Schema markup for AI: what it does for AI answers

Schema markup does not get you cited in AI answers by itself. No major AI engine has said it uses structured data to choose which sources to quote, and Google states that its AI Overviews and AI Mode need no special markup. What schema does is describe your page in a form machines read without guessing: who you are, what the product costs, who wrote the article and when. That helps search engines, which feed most AI answers, and it never hurts. Treat it as hygiene, not as a lever.

What structured data does for AI answers

Most AI answers start with a web search. ChatGPT, Gemini, Perplexity and Google's AI features read pages that rank in a search index, then write from what they read. Structured data touches that process in a few places:

• Search results. Google uses schema for rich results (product prices, review stars, breadcrumbs). Rich results can help a page earn clicks and links, which helps it rank, which is what gets it read.
• Entity facts. Organization markup with your name, logo and sameAs links to your profiles makes it easier to connect your site to your brand.
• Unambiguous facts. A price, a date or an author in markup matches the visible text and removes a chance of misreading.

What it does not do

• It does not guarantee a citation. An engine quotes the page that answers the question best, whatever its markup.
• It does not replace visible text. Facts that exist only in JSON-LD and not on the page are a mismatch, and Google's guidelines say markup must describe visible content.
• It does not fix access. If robots.txt blocks an AI crawler, or the content only appears after JavaScript runs, markup will not help.
• FAQ and HowTo rich results are mostly gone. Google limited FAQ rich results to a small set of sites and dropped HowTo rich results in 2023. Adding FAQPage for AI reasons alone is not worth much.

Schema types worth adding

Use JSON-LD in a <script type="application/ld+json"> block. It is the format Google recommends and the easiest to keep in sync with templates.

How to check your markup

1. Google's Rich Results Test shows which rich result types a URL qualifies for and any errors.
2. The Schema Markup Validator (validator.schema.org) checks any schema.org markup, including types Google does not use for rich results.
3. Search Console reports structured data errors across the site under the Enhancements section, once Google has crawled the pages.
4. View the raw HTML (not the browser inspector) and confirm the JSON-LD is in the server response. Some AI crawlers do not run JavaScript, so markup injected by a tag manager may never reach them.

zaDojo's site audit reads each crawled page's HTML and records whether it has a JSON-LD block. It does not validate the markup, so use the tools above for errors.

Where to spend the time instead

If you have an hour for AI visibility, schema is rarely the best use of it. In order:

1. Find out which pages AI answers already cite for your buying questions. See how to track your brand in ChatGPT.
2. Make your own pages answer the question in the first paragraph, with current prices and a comparison table. The GEO guide has a checklist.
3. Get onto the third-party lists and reviews those answers cite. The AEO guide covers how to find and pitch them.
4. Then add or fix Organization, Article and Product markup so your facts are consistent everywhere.

For Google specifically, see AI Overview optimization.

Questions people ask

Does ChatGPT read schema markup? OpenAI has not said how its search uses structured data. ChatGPT search reads pages found through search providers and its own crawler, so markup in the HTML can be read, but there is no evidence it decides citations.

Should I add FAQ schema for AI Overviews? Google says AI Overviews need no special markup. A clear question and answer in the visible text matters more than the markup around it.

Is llms.txt a kind of schema? No. It is a plain text file that points agents to your important pages. See the llms.txt guide.