Schema.org types Google AI Overviews actually surfaces — and the ones it ignores

Google AI Overviews — the generative summary panels that appear above organic Search results — surface content from the same Search index as the regular results. But the panel's generation leans heavily on structured data to identify entities, attribute facts, and decide which sources to cite. Pages with rich Schema.org markup are reliably overrepresented in AI Overview citations relative to their organic rank.

Below are the Schema.org types most worth shipping if AI Overview presence matters, ranked by how often Google's own structured data documentation references them as eligible for AI/answer features.

Tier 1 — almost universally worth adding

Organization (or LocalBusiness) — Establishes the entity that owns the site. Without an Organization block, AI tools have to infer from page content; with one, attribution is unambiguous. name, url, logo, sameAs (links to social/Wikipedia) are the high-leverage fields. LocalBusiness is preferred for physical-presence businesses because it carries address, telephone, openingHours.

Article / NewsArticle / BlogPosting — Tags editorial content as a citable article with headline, datePublished, author, publisher. AI Overviews preferentially cite content with explicit datePublished because it lets the model reason about freshness.

BreadcrumbList — Helps AI understand site hierarchy. Pages buried under a clear breadcrumb path (Home > Category > Article) get more reliable category attribution than orphan pages.

Tier 2 — high-impact for specific verticals

FAQPage — Direct fit for question-answer surfaces. AI Overviews routinely cite FAQ markup when the user query maps onto an answered question. The Question / Answer pair structure is exactly what generative models want to consume.

HowTo — Procedural content. Step-by-step markup is directly extractable into the kind of numbered list AI Overviews produce.

Product — For e-commerce and SaaS pages. Includes aggregateRating, offers, review. AI Overviews answering "best X for Y" queries lean heavily on Product markup with aggregateRating because it gives them a defensible signal to rank by.

Recipe — Vertical-specific but exemplary: AI Overviews almost always cite Recipe-marked content for cooking queries.

Event — For event listings. startDate, location, offers are surfaced.

Course — For educational content. Often surfaced for "how to learn X" queries.

Tier 3 — useful but secondary

Person, Place, Review, VideoObject, Dataset, JobPosting. Each has a documented Search appearance feature; whether AI Overviews surface them depends on query intent.

What AI Overviews appear NOT to weight

A few popular Schema.org types don't seem to move AI Overview surfacing despite being valid markup:

• WebPage, WebSite — too generic to be useful for entity attribution beyond what the URL already conveys.
• CreativeWork — too abstract; prefer the more specific subtype (Article, Book, etc).
• Custom types not in the Schema.org spec — Google's documented validators don't surface unknown types in Search appearance features.

Implementation notes

JSON-LD in a <script type="application/ld+json"> tag is the recommended format — Google's docs explicitly prefer it over Microdata or RDFa. The Rich Results Test at https://search.google.com/test/rich-results validates whether your markup is eligible for any Search appearance feature.

For AI Overview specifically there is no public "rich result" type to validate against — Google has not published a list of "AI Overview eligible" markup. The signal we observe in production probes is that markup eligible for any Search appearance feature (FAQ, Article, Product, etc.) correlates with higher AI Overview citation rate.

What AI Visibility Monitor actually checks today

Honest scope as of methodology v2.0:

Currently checked by the Entity Clarity scorer (in the AI Visibility Analyzer, linked from Monitor's Readiness page):

• Organization schema presence
• Organization completeness — name, url, logo, description, sameAs, contactPoint
• Author schema on content pages

That's the foundational entity layer — necessary but not sufficient for AI Overview eligibility.

Not yet checked (tracked as planned coverage):

• LocalBusiness (preferred for physical-presence businesses)
• Article / NewsArticle / BlogPosting
• BreadcrumbList
• FAQPage
• HowTo
• Product with aggregateRating

These are the higher-impact AI Overview signals identified in the research above. We score Organization/Author today because that's the most universal foundation; the verticalised types (FAQ, Product, Article) ship in a later release. When they do, the methodology version bumps and a follow-up research piece will document what changed and why.

If you've added these types to your site already, the Rich Results Test at https://search.google.com/test/rich-results validates them independently of our coverage.