Schema markup for AI search
Schema markup is structured data, usually written as JSON-LD, that labels what a page is about so engines do not have to guess. For AI search the most useful types are Organization and Person (who published this), Article (what it is and when), FAQPage and HowTo (liftable answers), and Product and Review (commerce). Put it in the server-rendered HTML so no-JS crawlers can read it.
- →Schema markup is structured data that tells an engine what a page is, using vocabulary from Schema.org, usually as JSON-LD.
- →The types that matter most for AI search are Organization, Person, Article, FAQPage, HowTo, Product and Review.
- →Schema must be in the server-rendered HTML, because ChatGPT, Perplexity and Claude do not run JavaScript.
What is schema markup?
Schema markup is structured data that describes a page in a vocabulary engines understand, defined by Schema.org and usually embedded as JSON-LD. Instead of leaving an engine to infer that a page is an article by a named author, or a product with a price and a rating, schema states it explicitly. This helps an engine resolve your entity, classify the content, and pull the right facts into an answer. Schema does not change what a reader sees; it is a machine-readable layer that sits alongside the visible content.
Which schema types help AI search?
A handful of Schema.org types do most of the work for AI search. Each answers a question an engine asks, and each maps onto one of Rubric's three pillars.
| Schema type | What it tells an engine | Pillar |
|---|---|---|
| Organization | The company or brand behind the page, with sameAs links | Known |
| Person | The author or named individual, with credentials and sameAs | Known |
| Article | That the page is an article, its headline, author and dates | Known and Trusted |
| FAQPage | A set of question-and-answer pairs an engine can lift | Findable |
| HowTo | Ordered steps for a task | Findable |
| Product | A product with attributes, price and availability | Known |
| Review / AggregateRating | A rating or review an engine can trust and quote | Trusted |
Add the types that match the page. A blog post wants Article plus Person; a support answer wants FAQPage; a task guide wants HowTo; a shop page wants Product and Review.
Why schema must be server-rendered
Schema only helps if the engine can read it, and several engines cannot read schema added by JavaScript. ChatGPT, Perplexity and Claude do not run JavaScript, so JSON-LD injected by a tag manager after the page loads is invisible to them, even though it looks correct in the browser inspector. Google renders JavaScript and Microsoft Copilot does so unreliably, so client-side schema is a gamble at best. Put JSON-LD in the server-rendered HTML, in the head, so it is present in the raw response. See why-ai-crawlers-dont-run-javascript for the render-parity detail and the View Source test.
What good Organization schema looks like
The single highest-leverage schema on most sites is Organization with sameAs, placed in the shared head so one change covers every page. Here is a minimal, valid example.
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "GoGoChimp",
"url": "https://gogochimp.com",
"logo": "https://gogochimp.com/logo.png",
"sameAs": [
"https://www.linkedin.com/company/gogochimp",
"https://www.wikidata.org/wiki/Q139585936"
]
}The sameAs array is what lets an engine connect your site to a known entity rather than guessing. See entity-clarity-and-sameas for how to choose those links, and adding-organization-schema for the step-by-step.
Does schema guarantee a citation?
No. Schema does not force an engine to cite you; it removes ambiguity so the engine understands your page and can resolve who published it. A citation still needs a liftable answer near the top (Findable) and evidence the engine trusts (Trusted). Schema mainly strengthens the Known pillar, and on its own it will not carry a page that has no clear answer or no sourced evidence.
How Rubric checks schema
Rubric scores schema on more than its presence. It checks that content-type schema is present for the page, that the schema is complete and valid rather than half-filled or malformed, and, crucially, that it is readable without JavaScript. Organization and Person schema sit under the Known pillar; FAQPage and HowTo sit under Findable; review and rating schema sit under Trusted. A page can have schema and still fail the check if the markup is invalid or only appears after scripts run.