Schema Markup for AI Search: What It Does, What Google Says, and What to Add

What schema markup is

Schema markup is structured data added to a page, usually as a block of JSON-LD, that labels the facts already on it using the shared Schema.org vocabulary: this is an organization, this is its name, this is its product, this is the price. Search engines use it to understand a page with less guesswork, and Google uses some types to show rich results such as product details and breadcrumbs.

For answer engine optimization, schema matters because AI engines draw on the same understanding of entities that search engines build. Our Knowledge Base entry on structured data for GEO covers the concept; this guide covers what the evidence says and what to implement.

What Google and Microsoft have said about schema and AI

The two companies have said different things, and both statements are worth reading carefully.

Google Search's guide to optimizing for generative AI features, published in May 2026, says structured data isn't required for generative AI search and that you don't need special Schema.org markup. It puts special markup alongside llms.txt files and content chunking among the things you don't need. That does not make structured data useless on Google. The same guide recommends keeping it for rich results, and Google's structured data documentation still explains how markup helps it understand pages. The guide's point is narrower: there is no special markup that unlocks AI Overviews or AI Mode.

Microsoft has been more positive. At SMX Munich in March 2025, Bing's Fabrice Canel said schema markup helps Microsoft's LLMs understand content, as Search Engine Land reported. In the same context, he advised publishers to push fresh content through IndexNow, because generative AI systems tend to value freshness.

Does schema help you get cited in AI answers?

No engine has said that schema earns citations, and no published study has shown that adding schema alone changes how often a brand is cited in AI answers.

The indirect case is stronger, and it is enough to justify the work:

• Clear entities. Organization markup with links to your official profiles helps systems connect your site, your LinkedIn page, your review profiles and your mentions as one company.
• Accurate facts. Prices, plans and product details in markup give machines a clean version of the facts AI answers most often get wrong.
• Consistency. Markup that matches your visible content and your profiles reinforces one version of your company across the web.
• Rich results. Classic search still drives discovery, and rich results still depend on structured data.

What schema cannot do is make a page worth citing. A citation still depends on whether the page answers the prompt better than the alternatives.

Which schema types matter for B2B SaaS

Start with the types that describe who you are and what you sell. Most SaaS sites need only a handful.

Schema type	What it tells machines	Who should use it	Priority
Organization	Your name, logo, URL and official profiles (through sameAs)	Every company site	High
SoftwareApplication or Product	What your product is, what it does, which platforms it runs on	SaaS and software companies	High
Offer	Prices and plans	Sites with public pricing	Medium
Article with Person	Who wrote a piece, their role, and when it was published or updated	Blogs and resource libraries	Medium
BreadcrumbList	Where a page sits in your site structure	Larger sites	Low
FAQPage	Questions and answers on a page	Use sparingly: FAQ rich results are limited	Low

One change catches many teams out. In August 2023, Google limited FAQ rich results to well-known, authoritative government and health websites and stopped showing HowTo rich results. FAQ markup is still valid, but for most B2B sites it no longer earns a visual result, so it should not be the first type on your list.

How to implement and validate schema

1. Use JSON-LD. It is the format Google recommends, and it keeps markup separate from your page design.
2. Mark up what is visible. Google's guidelines require structured data to describe content users can see on the page. Never let a fact exist only in markup.
3. Render it on the server. Vercel and MERJ's analysis of AI crawler traffic (December 2024) found that none of the major AI crawlers, including GPTBot and ClaudeBot, render JavaScript. Markup injected by client-side scripts, for example through a tag manager, may never reach them. Put it in the HTML your server sends.
4. Connect your entities. Use sameAs on Organization to link your official profiles, and keep names, URLs and logos identical everywhere.
5. Validate it. Run Google's Rich Results Test for rich result eligibility and the Schema Markup Validator for general correctness.

A minimal Organization example, for a fictional company:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Acme Analytics",
  "url": "https://acme.com",
  "logo": "https://acme.com/logo.png",
  "sameAs": [
    "https://www.linkedin.com/company/acme",
    "https://www.g2.com/products/acme"
  ]
}

Entity work goes beyond schema. Our Knowledge Base entry on entity and knowledge graph optimization explains how consistent entity information helps engines recognize your brand.

Common schema mistakes

• Markup that does not match the page. It breaks Google's guidelines and teaches machines a version of your facts no reader can verify.
• Outdated prices and features. Stale Offer or Product markup feeds the same errors you are trying to prevent.
• Invented "AI" properties. There is no special schema type for AI search; made-up properties are simply ignored.
• FAQ markup used as marketing copy. Questions written to stuff keywords help nobody and no longer earn rich results on most sites. If you add FAQ markup, generate it from the questions the page actually answers, for example with our free FAQ schema generator.
• Facts that disagree across sources. If your schema, website and profiles give different company details, every system reading them has to guess.

Where schema fits with llms.txt and crawler access

Schema is one of three technical items that appear on every AI search checklist, and each does a different job.

Item	What it does	What it cannot do	Priority
robots.txt	Decides which crawlers may fetch your pages	Make engines cite you	Check first: a blocked crawler sees nothing
llms.txt	Points AI tools to your key pages	Allow or block anything	Low for most marketing sites
Schema markup	Labels the facts on each page	Earn citations on its own	Medium: worth doing well once

For the details on the other two, see our guide to what llms.txt does and doesn't do and the crawler access checks in our AEO audit guide.

Omnia's view

Fix the basics once: Organization, your product, pricing and authorship, server-rendered, valid and identical to what your pages say. Then put your effort where AI answers are decided: pages that answer buyers' questions clearly, facts that match everywhere they appear, and mentions on the sites engines already trust.

Schema makes your facts easier to read. It cannot make them worth citing. If you want to see which of your pages AI engines cite today, and which sources they pick instead, that is what Omnia tracks, and what the Omnia Agent turns into work your team can ship. For the wider approach, see our guides to answer engine optimization and to earning visibility in AI Overviews.

Start your free 14-day trial of Omnia.