What is Schema Markup?
schema.org structured data (JSON-LD) that tells AI engines what your pages and brand are.
What schema markup is
Schema markup is machine-readable data added to a web page that states what the page is, in a vocabulary search engines and AI answer engines have agreed to interpret consistently. The vocabulary comes from Schema.org, maintained collaboratively by Google, Microsoft, Yahoo, and Yandex. On the web today it is nearly always embedded as JSON-LD - a block of JSON inside a script tag in the page's <head>.
Typical use cases include declaring an article's author and publish date (Article), a product's price and availability (Product / Offer), a set of FAQ Q/As (FAQPage), and a company's name and identity (Organization). Rich search results and answer-engine attribution both rely on these signals to remove ambiguity about what the page contains.
What schema markup does and does not do
It removes ambiguity about the page's contents and its publisher. That helps rich-result eligibility on Google (star ratings, breadcrumbs, event dates) and gives AI engines a reliable way to attach a claim to a specific author or organisation instead of inferring both from layout.
It does not directly boost rankings. Google is explicit that schema markup is not a ranking signal. Marking up content that is not visible on the page is a policy violation, not a shortcut - if the FAQPage schema lists a question the reader cannot see, Google can drop the schema and issue a manual action.
Also known as: structured data, JSON-LD.
Frequently asked questions
- Does schema markup help me get cited by ChatGPT?
- Indirectly. Schema is not a confirmed citation lever for ChatGPT itself, but the content it clarifies (author, publisher, publish date, article type) is exactly what the retrieval and re-ranking layers use to decide whether a claim can be safely attributed. Sites cited most often in AI answers overwhelmingly carry basic Article and Organization schema, and Bing (the primary index behind ChatGPT Search) uses it heavily.
- What schema types matter most for AI citation?
- Organization on every page (to establish the publisher entity), Article on long-form content with datePublished / dateModified / author populated, FAQPage where you have real customer questions with answer-first prose, and Product / SoftwareApplication + Offer where you sell software. Every marked-up field must correspond to something the reader can see on the page.
- JSON-LD, Microdata, or RDFa - which format?
- JSON-LD is what Google, Bing, and every major AI engine recommend today. It sits in a script tag in the <head>, does not interleave with visible HTML, and is trivial to keep in sync with the source of truth (a shared TypeScript module, in Osoro GEO's case). Microdata and RDFa still work but are strictly harder to maintain.
Every Osoro GEO scan measures this and the rest of the CITE framework on your own site - free, in about 30 seconds. How the CITE framework works →