Schema markup for AI crawlers: what to add and why

In traditional SEO, schema markup earns rich snippets — star ratings, FAQ dropdowns, breadcrumbs. Useful, but supplementary. For AI crawlers, schema serves a deeper purpose: it provides machine-readable context that the crawler would otherwise have to infer from raw HTML. When GPTBot encounters a page with Article schema, it knows the headline, author, publish date, and modified date without parsing the visual layout. When it finds FAQPage schema, it has clean question-answer pairs ready to cite. According to a 2024 Schema.org adoption study, pages with structured data receive 35–40% more clicks from rich results — and early evidence suggests similar or greater advantages in AI citation rates.

This matters because AI crawlers are making trust decisions at scale. They process millions of pages and need to quickly determine: who wrote this, when was it updated, what entity stands behind it, and is this a direct answer to a question? Schema answers all four questions in structured, unambiguous format. Microsoft’s Copilot documentation explicitly recommends making product catalogues and content ‘machine-readable’ — JSON-LD structured data is the most direct path to that goal.

Not all schema types are equally useful for AI citation. Focus on these five, in priority order.

Always use JSON-LD format embedded in a <script type='application/ld+json'> tag. Microdata and RDFa are technically valid but harder for AI crawlers to parse reliably, and Google has explicitly recommended JSON-LD for years.

Use a @graph array to combine multiple schema types on one page — for example, Article + FAQPage + BreadcrumbList on a resource article. Keep the @id references consistent so engines understand the relationships between entities. Test with Google's Rich Results Test and Schema Markup Validator before deploying.

The most damaging mistake is fabricated schema — adding review ratings you don't have, author credentials that don't exist, or product specs that don't match reality. AI engines cross-reference schema against page content and external sources. If the schema claims 4.8 stars but the page shows no reviews, the engine trusts you less, not more.

Other common mistakes: using outdated dates (schema says 2024 but the content references 2026 events), missing author information (anonymous content is harder to trust), and orphaned schema (Organization schema on a single page instead of sitewide). Fix these before adding more schema types.