Structured data in AI answers
What Google, ChatGPT and Perplexity actually document about markup and page structure, separated from guesswork, with practical guides for each part.
- Guides
- 8
- Types covered
ArticleProductOrganization+13- Sources
- Google Search Central, schema.org
- Updated

What we know vs what is claimed
LLM SEO, also sold as AEO or GEO, is the practice of shaping pages so that AI answer engines understand them and cite them. Much of the advice in that market is confident and unsourced. This section of the handbook takes the opposite approach: every guide starts from what Google Search Central, the ChatGPT crawler documentation and the Perplexity crawler documentation actually say, and marks plainly where the documentation stops.
The central finding is easy to state. Google's page on AI features and your website says there is no special schema.org structured data you need to add to appear in AI Overviews or AI Mode. Structured data still helps Google understand pages and keeps them eligible for rich results. The guide Does schema markup help AI Overviews? works through that position in detail, including why no public controlled test settles the question.
How each engine fetches pages
Before any markup matters, a crawler has to reach the page. Google's AI features use pages from its Search index. ChatGPT search uses OAI-SearchBot, Perplexity uses PerplexityBot, and both send separate fetchers when a user asks a question. Neither of the last two documents whether it renders JavaScript or reads JSON-LD. The guide to how ChatGPT and Perplexity read your pages lists each user agent, what robots.txt controls, and a simple test you can run from your own logs.
Markup that helps understanding
Some markup describes things that are hard to infer from prose: who publishes a site, who wrote an article, when it last changed. The guide to Organization schema and entity markup shows how to describe the publisher once with an @id and confirm it with sameAs links. The guide to Author and Person schema does the same for writers, and the guide to Article schema and dateModified explains how to keep dates honest and equal to what the page shows. If you are new to the vocabulary itself, start with what is structured data? on the home page.
Trust signals
AI answers make thin or fake signals easier to spot, and the documentation is moving the same way. Google's review snippet guidelines now name fake and undisclosed incentivized reviews explicitly, and its helpful content guidance treats fabricated author profiles as deception. The guide to review ratings in AI answers explains where star ratings in AI answers come from, which is mostly product feeds and business profiles rather than page markup alone, and why honest reviews matter more than stars.
Quotable structure
The part of a page that gets quoted is text. A direct answer in the first sentence under a heading, comparisons in a captioned HTML table and one-sentence definitions make a passage usable on its own. The guide to getting cited by ChatGPT and AI answers turns that into a checklist and covers the snippet controls, such as nosnippet and data-nosnippet, that can block quoting by accident.
One former favourite has dropped out of Google's results entirely. FAQ rich results were limited to government and health sites in 2023 and stopped appearing in May 2026. The guide to FAQ and HowTo schema after the cuts gives the timeline from Google's own announcements and a plain answer on whether to keep the markup.
How to choose a guide
If you have one question, pick the guide that matches it:
- Wondering whether markup is worth the effort for AI search: read the guide on schema and AI Overviews first.
- Checking why a crawler is not visiting: read the crawler and fetcher guide.
- Cleaning up a WordPress theme that prints messy author or publisher markup: read the Person guide, then the Organization guide.
- Rewriting a page so it can be quoted: use the quotable page checklist.
How to measure what happens
Google reports AI feature traffic inside Search Console rather than in a separate tool. Its guide to generative AI features points site owners to a Generative AI performance report, and warns that no third-party tool has access to Google's internal ranking or AI systems. For ChatGPT and Perplexity, server logs showing visits from their crawlers, and referral visits from their answer pages, are the most direct evidence available to a site owner.
Every guide states the date its sources were read, links each claim to the page that supports it, and says so when something is not documented. No guide promises rankings, stars or citations, because no engine does.
Every guide in this section
Register · 8 entries| Guide | What it answers | Schema type | Updated |
|---|---|---|---|
| Does schema markup help AI Overviews? | What Google says about schema and AI Overviews, and where markup still pays | Article Product | |
| How ChatGPT and Perplexity read your pages | Crawlers, raw HTML and JSON-LD: what ChatGPT and Perplexity are documented to read | Article WebPage | |
| How to get cited by ChatGPT and AI answers | Direct answers, tables, headings and snippet controls: a quotable page checklist | WebPage SpeakableSpecification | |
| Author and Person schema | Person markup that ties an article to a real, checkable writer | Person Article | |
| Organization schema and entity markup | Describe your site as one clear entity with Organization, @id and sameAs | Organization WebSite | |
| Article schema and dateModified | datePublished, dateModified and honest freshness for articles | Article BlogPosting | |
| Review ratings in AI answers | Where AI answers get star ratings, and what on-page review markup still does | Review AggregateRating | |
| FAQ and HowTo schema after the cuts | What happened to FAQ and HowTo rich results, and whether to keep the markup | FAQPage HowTo |
The guides
Three drawersHow AI engines read pages 3 guides
- AI answersCERT No. AI 0001

Does schema markup help AI Overviews?
What Google says about schema and AI Overviews, and where markup still pays
- Type
- Article, Product
- Updated
- AI answersCERT No. AI 0002

How ChatGPT and Perplexity read your pages
Crawlers, raw HTML and JSON-LD: what ChatGPT and Perplexity are documented to read
- Type
- Article, WebPage
- Updated
- AI answersCERT No. AI 0008

How to get cited by ChatGPT and AI answers
Direct answers, tables, headings and snippet controls: a quotable page checklist
- Type
- WebPage, SpeakableSpecification
- Updated
Entity and trust signals 3 guides
- AI answersCERT No. AI 0004

Author and Person schema
Person markup that ties an article to a real, checkable writer
- Type
- Person, Article
- Updated
- AI answersCERT No. AI 0005

Organization schema and entity markup
Describe your site as one clear entity with Organization, @id and sameAs
- Type
- Organization, WebSite
- Updated
- AI answersCERT No. AI 0006

Article schema and dateModified
datePublished, dateModified and honest freshness for articles
- Type
- Article, BlogPosting
- Updated
Rich results that changed 2 guides
- AI answersCERT No. AI 0003

Review ratings in AI answers
Where AI answers get star ratings, and what on-page review markup still does
- Type
- Review, AggregateRating
- Updated
- AI answersCERT No. AI 0007

FAQ and HowTo schema after the cuts
What happened to FAQ and HowTo rich results, and whether to keep the markup
- Type
- FAQPage, HowTo
- Updated
The other three sections
Elsewhere in the handbook- Section STSchema typesHow each schema.org type works today, property by property, and what it can still earn.
- Section WPWordPressExact steps for adding, testing and fixing markup in WordPress and WooCommerce.
- Section AUAuthenticityHonest reviews, disclosed AI assistance and provenance labels for images, in markup.