$ man geo/schema-markup-for-ai

Schema markup is the bridge between your content and AI engines' structured understanding of it. When you add JSON-LD schema to your pages, you are giving AI engines a machine-readable layer of metadata that tells them exactly what your content is, who created it, when it was updated, and how its parts relate to each other. Without schema, AI engines have to infer all of this from your HTML, which is ambiguous. A date on your page could be the publication date, a date mentioned in the content, or a date in a sidebar widget. Schema removes that ambiguity by explicitly declaring: this is the datePublished, this is the dateModified, this is the author, this is the article body. AI engines use schema in two ways. First, they use it for content understanding - identifying what the page is about and what claims it makes. Second, they use it for trust signals - author credentials, publication dates, and organizational affiliations all feed into authority evaluation. Pages with comprehensive schema markup consistently outperform pages without it in AI citation rates.

PATTERN

Focus on these schema types for maximum GEO impact. Article schema: use this on every content page. Include headline, description, author (as a Person with url and sameAs pointing to social profiles), datePublished, dateModified, publisher (as an Organization), image, and mainEntityOfPage. This is the foundation. FAQPage schema: use this on pages with question-and-answer content. Each question-answer pair becomes a separate extraction opportunity for AI engines. This schema directly maps to how AI engines process FAQ content. HowTo schema: use on tutorial and guide pages. Include step-by-step instructions with name, text, and optional image for each step. AI engines use this to generate step-by-step answers. Organization schema: use on your about page and homepage. Include name, url, logo, sameAs (array of social profile URLs), foundingDate, and description. This is how you declare your entity to AI engines. Person schema: use for author pages and about sections for individual contributors, linking to their credentials and social profiles.

CODE

In a Next.js or React-based content system, implement schema as a component that generates JSON-LD in the page head. Create a reusable SchemaMarkup component that accepts a type and data props, then renders a script tag with type application/ld+json containing the serialized schema object. For the ShawnOS.ai monorepo, schema is generated at the page level in the wiki and blog templates. Each wiki entry automatically gets Article schema with the entry title as headline, the description as the schema description, the current date as dateModified (because wiki content is continuously updated), and the author as a Person entity with sameAs links. The key implementation detail: set the context to https://schema.org and the type to the appropriate schema type. Nest objects using the at-type property. Arrays of items - like FAQ questions or HowTo steps - are represented as arrays of typed objects. Validate your schema using Google's Rich Results Test or Schema.org's validator to catch errors before deploying.

PRO TIP

Of all the schema properties you can implement, dateModified has the highest impact on AI citations. AI engines use it as the definitive freshness signal for your page. If your schema says the page was modified yesterday, AI engines treat it as yesterday's content regardless of when the URL was first published. This makes dateModified a powerful lever. When you update a page, always update the dateModified in your schema. Automate this - in a Next.js system, you can set dateModified to the build date or the git commit timestamp for that file. Do not set dateModified to the current server time on every page load - that is spammy and AI engines will learn to ignore your freshness signals. Set it to the actual last meaningful update. And keep it consistent with your other freshness signals: your sitemap lastmod, your RSS feed dates, and your HTTP Last-Modified header should all align with your schema dateModified.