Schema Markup for AI Answers: The JSON-LD That Matters for B2B

In this article

1. What schema markup does (and what it doesn't)
2. What the platforms and the 2026 data say
3. The five schema types a B2B site needs
4. Organization: the one that matters most
5. Service and Product markup
6. Article and FAQPage markup
7. Implementation rules that matter for AI
8. Where schema sits in your AI visibility priorities
9. Frequently asked questions

What schema markup does (and what it doesn't)

Schema markup is structured data, usually written as JSON-LD, that labels the facts on a page using the shared vocabulary at schema.org. Instead of a machine guessing that "Toledo, Ohio" is your headquarters, the markup says so explicitly: this is an Organization, this is its address, this is its logo, and these are its official profiles.

For traditional search, schema powers rich results such as breadcrumbs, product details and review stars. For AI search, the claims run well ahead of the evidence. So let's start with what's actually known.

What the platforms and the 2026 data say

• Google: its guide to generative AI features says "Structured data isn't required for generative AI search" and "there's no special schema.org markup you need to add." It recommends structured data for rich results as part of normal SEO.
• Ahrefs (May 2026): 1,885 pages that added JSON-LD between August 2025 and March 2026 were compared with matched control pages. ChatGPT citations changed by +2.2% and AI Mode by +2.4%, neither statistically different from zero. AI Overview citations fell slightly (−4.6%). The conclusion: "Adding schema produced no major uplift in citations on any platform." The study only covered pages that were already heavily cited, so it can't say whether schema helps new pages get noticed.
• Microsoft: Bing's Fabrice Canel was reported to have confirmed at SMX Munich in March 2025 that schema helps Microsoft's language models understand content for Copilot. The report was second-hand, from an attendee's post, but it's the closest thing to an official statement from any AI platform.
• OpenAI, Anthropic and Perplexity haven't published anything about how they use schema.

The five schema types a B2B site needs

Companies with physical sites customers visit should add LocalBusiness (or a more specific subtype) for each location. Skip HowTo: Google retired HowTo rich results in 2023.

Organization: the one that matters most

Organization markup is your machine-readable company profile. Give it a stable @id so other markup on the site can point to it, and list your official profiles in sameAs. That's how machines connect your website to the same company on LinkedIn, Crunchbase, Wikidata and industry directories.

Organization JSON-LD for a fictional distributor

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Organization",
  "@id": "https://www.example.com/#organization",
  "name": "Northfield Industrial Supply",
  "legalName": "Northfield Industrial Supply, Inc.",
  "url": "https://www.example.com/",
  "logo": "https://www.example.com/logo.png",
  "description": "Industrial MRO distributor serving manufacturers in Ohio, Michigan and Indiana.",
  "foundingDate": "1987",
  "numberOfEmployees": { "@type": "QuantitativeValue", "value": 240 },
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "100 Example Ave",
    "addressLocality": "Toledo",
    "addressRegion": "OH",
    "postalCode": "43604",
    "addressCountry": "US"
  },
  "contactPoint": {
    "@type": "ContactPoint",
    "contactType": "sales",
    "telephone": "+1-419-555-0100",
    "email": "[email protected]",
    "areaServed": "US"
  },
  "areaServed": ["Ohio", "Michigan", "Indiana"],
  "sameAs": [
    "https://www.linkedin.com/company/example",
    "https://www.crunchbase.com/organization/example"
  ]
}
</script>

Only include facts that are true and visible somewhere on your site. If your website says 240 employees and LinkedIn says 300, fix the mismatch first. Conflicting facts are exactly what makes AI hedge.

Service and Product markup

On each service page, describe the service and point back to your Organization by its @id. This ties the offering to the company without repeating every detail.

Service JSON-LD linked to the Organization

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Service",
  "name": "On-site PPE vending programs",
  "serviceType": "Industrial vending and inventory management",
  "description": "On-site vending machines for PPE and consumables, restocked weekly, with usage reports by department.",
  "provider": { "@id": "https://www.example.com/#organization" },
  "areaServed": ["Ohio", "Michigan", "Indiana"],
  "url": "https://www.example.com/vending"
}
</script>

Use Product instead for physical products with SKUs. Add offers with price only if the price is public on the page, since markup must match what visitors see.

Article and FAQPage markup

For blog posts and guides, Article or BlogPosting markup should include a named author (a Person with a URL), the publisher (your Organization's @id), and datePublished and dateModified. Named expertise and visible update dates help readers and machines judge how current and credible a page is. Every guide on this site uses this pattern. View the source of this page to see it.

Use FAQPage only for a real, visible FAQ. Since 2023 Google shows FAQ rich results only for well-known government and health sites, so the benefit is clarity, not a bigger search listing.

Implementation rules that matter for AI

1. Put it in the HTML, not in Tag Manager. Google renders JavaScript and can read JSON-LD injected by GTM. Most AI crawlers can't. See why JavaScript content is invisible to AI.
2. Connect your entities with @id. One Organization, referenced by your services, articles and people, instead of repeated, slightly different copies.
3. Match the visible page. Google's structured data guidelines require markup to describe content users can see. Hidden claims risk penalties and contradict your own page.
4. Keep facts consistent everywhere: schema, page copy, llms.txt, LinkedIn and directories.
5. Validate it. Use Google's Rich Results Test for Google features and the Schema Markup Validator (validator.schema.org) for general syntax.
6. Don't mark up self-serving reviews. Review stars for your own business on your own site aren't eligible in Google, and fake ratings are a policy violation.

Where schema sits in your AI visibility priorities

If you're deciding where to spend the next month, schema is not first. In order: make sure AI crawlers can reach you (robots.txt and firewall), make sure they can read you (server-rendered HTML), make your key pages state facts plainly, then add the five schema types above. After that, earn mentions on the third-party sites AI assistants cite. Schema is the fourth step, not the first.

Frequently asked questions

Free tool

Does your homepage have Organization schema?

The free LLM Readiness Check looks for JSON-LD structured data such as Organization, along with crawler access, raw HTML, sitemap and llms.txt.

Run the free check