Schema for AI search: every schema type that actually moves citation rate
Schema markup is the structured data layer AI engines extract from to answer questions. If you’re trying to figure out which schema types matter for AI search citation, which ones are nice-to-haves, and which ones are wastes of time, this is the hub.
Every post we’ve written on schema sits below. Read in order if you’re starting from zero, or skip to the type you’re shipping this week.
Where to start if you’re new to schema for AI search
Read Schema markup for AI search: a 2026 field guide first. It’s the overview of the category and explains why schema matters more for AI engines than it did for classic Google SEO.
Then read JSON-LD vs Microdata vs RDFa in 2026 for the format decision. Short version: JSON-LD wins, ship it everywhere.
Then come back here and pick the type you need.
The schema types that matter most (in order of citation impact)
Organization schema
Your Organization schema is the foundation. Without it, AI engines can’t confidently route citations to your domain because there’s no canonical entity. Everything else builds on top.
The fields that matter most are sameAs (pointing to LinkedIn, X, Crunchbase, Wikidata), founder (binding to your Person schema via @id), and knowsAbout (5-8 specific topics your brand has authority on).
Read:
– Your Organization schema is probably costing you AI citations
FAQPage schema
FAQPage schema is the single highest-extraction format AI engines use. Pages with FAQPage schema get cited at 2-3x the rate of identical content without it. If you ship one schema type this quarter, ship this one.
Read:
– FAQPage schema is the single biggest AI citation move
– Why FAQ content is the most underrated B2B content type in 2026
Person schema (for founder-led brands)
If your brand has a founder doing the work or being the public face, Person schema binds that founder entity to the brand entity. Article schema’s author field references the Person via @id. The whole entity graph compounds across every blog post.
Read:
– Person schema for founder-led brands (the schema everyone skips)
Product / SoftwareApplication schema
For SaaS companies, the pricing page is one of the highest-AI-query surfaces on your site. Product or SoftwareApplication schema with Offer + priceSpecification is what makes pricing extractable as structured data instead of guessed from prose.
Read:
– Product schema for SaaS: turning your pricing page into a citable answer
Article schema
Article schema is what makes blog content show up as editorial content rather than generic page content. The fields that matter are author (binding to Person via @id), publisher (binding to Organization), datePublished, dateModified, and headline.
Read:
– Article schema for AI search: when it matters, when it doesn’t
HowTo schema
HowTo schema turns tactical posts into structured procedures AI engines can extract step-by-step. The “how to set up X” or “how to ship Y” content type benefits most.
Read:
– HowTo schema mistakes that break the rich result
ItemList schema
ItemList schema is what powers comparison and ranking content. If you have /vs/ pages or “best X for Y” listicles, ItemList schema with structured items is what AI engines preferentially cite when prospects ask comparison questions.
Read:
– ItemList schema for /vs/ pages (the comparison signal we missed)
Speakable schema
Speakable schema marks parts of a page as suitable for voice assistants. The use case is narrower than other schema types (only certain content benefits), but for the right content, voice extraction goes from broken to clean.
Read:
– Speakable schema: the underused signal that helps voice assistants find your content
LocalBusiness schema (for service-area businesses)
Standard LocalBusiness schema fits storefront businesses. Service-area businesses (plumbers, electricians, mobile services, traveling specialists) need a different setup. Skip the address. Define areaServed and serviceArea. Use the specific subtype.
Read:
– Schema for service-area businesses: LocalBusiness done right
The schema types that don’t matter as much for AI search
A few schema types that get a lot of attention but don’t move citation rate the way the above types do.
WebSite and BreadcrumbList. Yoast handles these automatically. They’re useful for Google’s rich results but don’t meaningfully affect AI engine citation.
Event schema. Useful if you’re publishing events, but only relevant for events-focused brands. Most B2B SaaS doesn’t need it.
Recipe schema. Specific to food content. Not relevant for most.
VideoObject. Useful if you publish video, especially YouTube. Worth shipping if you do. Most B2B brands underuse it.
How to validate that your schema is actually working
Three validators we use on every schema deployment.
Google’s Rich Results test catches errors that would prevent rich result display.
Schema.org’s structured data validator catches type-level errors against the vocabulary.
Bing’s Webmaster Tools markup validator catches issues that affect ChatGPT browse and Microsoft Copilot specifically.
Beyond the standard validators, there are five additional checks we run for AI citation specifically (relationship @id graph audit, content consistency, external link verification, CSS selector existence, third-party platform cross-reference). Read about all of them in The schema validators most agencies use (and the issues they completely miss).
What we do at NetPageTwo
We ship schema across every client engagement based on a priority order tied to the client’s specific buyer and content surface. A typical engagement adds Organization + Person + sameAs in week one, FAQPage in week two, Article schema bound to Person via author in week three, and the specific commercial types (Product, Service, ItemList) in week four.
The deployment is one block of JSON-LD per page, shipped via either direct theme edit or a schema injection layer (Code Snippets plugin on WordPress, similar mechanism on other CMSs). We validate before shipping. We audit quarterly.
If you want a schema audit for your existing site, the fit call covers it. The audit ships in week one with prioritized recommendations.
Related hubs:
– AI engines
– AI content strategy
– GEO 101
– The new search
– Field Notes