Structured Data for Google AI Overviews: Which Schema Types Actually Matter

Structured data influences Google AI Overviews indirectly but materially: schema.org markup feeds the entity understanding and content classification that sit underneath passage retrieval, helping Google know what your page is, who published it, and which facts are canonical. It is not a citation switch — un-marked-up pages get cited constantly — but on competitive queries, clean JSON-LD is part of how equally relevant passages get separated.

How AI Overviews actually consumes markup

The overview pipeline retrieves passages from Google's ordinary index, where structured data has already done its work at indexing time: entity resolution (linking your Organization to its Knowledge Graph node), content typing (this page is a Product page, that one an Article), and fact grounding (prices, ratings, dates as typed values rather than scraped strings). When the Gemini layer composes an answer about your product's price, a typed offers.price value on the canonical page is much harder to garble than a number buried in styled markup. Schema, in other words, is anti-hallucination infrastructure more than a ranking lever.

The types ranked by AI Overview value

Deprioritize decorative types (WebPage boilerplate, SiteNavigationElement) — they add maintenance surface without answer-relevant facts.

A minimal high-value implementation

An Article block that pulls its weight looks like this:

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "How much does CRM software cost in 2026?",
  "author": { "@type": "Person", "name": "Jane Doe",
    "sameAs": "https://www.linkedin.com/in/janedoe" },
  "datePublished": "2026-01-10",
  "dateModified": "2026-07-05",
  "publisher": { "@type": "Organization", "name": "Acme" }
}

The named-author-with-sameAs pattern matters more than most properties: anonymous content is systematically weaker in Google's quality systems, and overviews inherit those judgments. Full type definitions live at schema.org.

The rich-results caveat that confuses everyone

In August 2023 Google restricted FAQ rich results to authoritative government and health sites and reduced HowTo rich results — a change about SERP display, not about markup consumption. The markup still parses, still types your content, and still supplies clean Q&A structure. Implement FAQPage for machine readability and skip any expectation of visual treatment. One hard rule survives from Google's guidelines: markup must match visible page content. Schema describing answers that are not on the page is a spam signal, not a shortcut.

Testing and maintenance

Validate every template three ways: Google's Rich Results Test (eligibility), the Schema.org validator (syntactic correctness), and a raw-HTML check confirming the JSON-LD ships server-side — schema injected by client-side JavaScript inherits all the rendering risks that afflict content itself. Then wire schema review into your content-refresh process: a dateModified that contradicts visible dates, or an offers.price that lags a pricing change, actively feeds wrong facts into answer composition. Menra's content AEO checks treat schema consistency as part of page-level answer readiness, and the glossary defines the structured-data vocabulary used here. Treat markup as a facts API you publish for Google — versioned, validated, and kept true.