Schema markup in 2026: what still earns rich results, and what AI reads instead
Google killed the FAQ rich result in May 2026. Here's the schema that still earns rich results, plus the structured data AI search actually reads.
If you added FAQ schema to your site sometime in the last few years hoping for the little expandable accordion under your listing, here's the update: it's gone. Google pulled the FAQ rich result from Search entirely as of May 7, 2026, following Search Console reporting for it in June and API support in August. The snippet you were chasing doesn't exist anymore, for anyone.
That's a real change, and it's a good excuse to ask a question most service businesses never get around to: what does schema markup actually do for you now? This is a cluster post in Content & AI Search, and it's the technical companion to the content pillar: the plumbing underneath the content, not the content itself.
What schema markup actually is
Schema markup is code, usually JSON-LD, sitting in a page's source that states facts explicitly instead of leaving them to be inferred. Your prose might say "we've been serving the valley for over twenty years and our customers love us." Schema says, in a format machines don't have to guess at: this is a LocalBusiness, this is the address, this is the phone number, this is the aggregate review rating, this is the price range. It's the difference between describing your business to someone and handing them your business card.
Prose describes your business to a reader. Schema states it to a machine. In 2026, you need both.
The FAQ rich result is dead. Read the fine print before you panic.
Google's own guidance is clear about the sequence: FAQ rich results were already restricted to "well-known, authoritative" government and health sites back in 2023, and as of May 2026 they're gone from Search entirely, for every site, no exceptions. If your FAQ section was earning that accordion snippet, it isn't anymore, and no amount of schema cleanup brings it back.
Here's the part worth sitting with, though: Google has also said it will keep using FAQ structured data to understand pages, even without rendering the visual snippet. And the FAQ pattern was never really about the snippet for AI purposes, it was about giving an answer engine a clean, quotable question-and-answer pair to lift. That value didn't disappear with the rich result. It just stopped being visible in classic blue-link search, which is exactly where AI Overviews and zero-click search were already pulling attention anyway.
What still earns a visible rich result
Losing FAQ doesn't mean schema stopped mattering. Several types still produce real, visible enhancements in search results, and they're the ones worth your time:
- LocalBusiness (or a specific subtype). This is what feeds knowledge panel details and reinforces the facts behind your Google Business Profile: name, address, hours, service area.
- Review and AggregateRating. Star ratings displayed next to your listing are one of the highest-leverage rich results available, and they tie directly into the review flow that already drives rankings and conversions.
- BreadcrumbList. A small enhancement, but a consistent one: it shows your site's structure directly in the result and follows Google's general structured data guidelines on nearly every page type.
- Organization and Person. These establish who's behind the content, which matters for E-E-A-T: a named author with real credentials, not an anonymous byline.
Pages carrying valid structured data see meaningfully higher click-through from the rich results they do earn, and the gap compounds across a whole site's worth of pages, not just one.
Why AI engines care about schema even when Google doesn't render it
Structured data has quietly become more important for a different audience: the AI systems deciding what to cite. ChatGPT, Perplexity, and Google's own AI Mode use structured data to verify claims and assess how trustworthy a source is before pulling from it. A page that states "4.9 stars, 1,800+ reviews" in AggregateRating markup is a fact an AI system can confirm and quote. The same number sitting only in a paragraph is a fact it has to infer, and inference is exactly where AI systems get cautious about citing at all.
This is the same logic behind everything we do in Generative Engine Optimization: the goal isn't to trick an algorithm, it's to make your facts unambiguous enough that a machine can repeat them confidently. Schema is the most literal version of that idea. It's also why every page on this site, including this one, carries JSON-LD schema in its source, whether or not a visitor ever notices.
Where to actually spend the hour
Don't chase every schema type schema.org documents; most don't apply to you and several, like FAQPage, no longer produce a visible result at all. Spend the time on LocalBusiness, Service, Review/AggregateRating, and BreadcrumbList, verified with Google's Rich Results Test so you know they validate, not just that they're present. That's the schema doing real work in 2026, for search engines and AI systems both.
If you're not sure whether your site's structured data is helping or just sitting there unvalidated, that's exactly the kind of technical gap the Growth Blueprint surfaces alongside the rest of your SEO and content audit. Structure precedes scale, down to the markup.