Structured Data for AI Assistants

Published: May 24, 2026

Structured data is a way to describe your content in a format that machines can read without guessing. Search engines have used it for years to generate rich results. AI assistants now use it too, and for a different reason: it tells them what your site actually is, not just what it looks like.

What structured data does for AI

When an AI assistant visits your homepage, it reads text, headings, and links. That gives it a rough picture. Structured data fills in the specific facts: your organization name, what you sell, your contact details, your product categories. Instead of inferring that you are a software company from your copy, the AI reads it directly from a machine-readable tag.

This matters because AI assistants are increasingly used to answer specific questions: "what does this company do", "is this tool free", "who made this product". Structured data is how you answer those questions accurately before the AI has to guess.

The format: JSON-LD

JSON-LD (JavaScript Object Notation for Linked Data) is the recommended format. It lives in a <script> tag in your page head and does not affect the visual layout of your page. Google, Bing, and AI crawlers all support it.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Acme Corp",
  "url": "https://acmecorp.com",
  "description": "Cloud-based project management for engineering teams.",
  "logo": "https://acmecorp.com/logo.png",
  "sameAs": [
    "https://github.com/acmecorp",
    "https://twitter.com/acmecorp"
  ]
}
</script>

Place this in the <head> of your homepage. One block is enough for most sites.

Choosing the right @type

The @type field tells AI assistants what category of entity you are. Pick the most specific one that fits:

  • Organization — general-purpose company or nonprofit
  • WebApplication — software you use in a browser (SaaS tools, apps)
  • LocalBusiness — physical location (restaurant, shop, clinic)
  • Product — a specific product with price and availability
  • SoftwareApplication — downloadable or installable software
  • Person — personal portfolio or professional profile

You can combine multiple types if needed. For example, a SaaS product can have both an Organization block describing the company and a WebApplication block describing the product.

Fields that matter most for AI

These fields have the highest impact on how AI assistants describe you:

  • name — your exact brand or product name. Keep it consistent with what appears on your site.
  • description — one to two sentences, factual, no marketing language. This is often cited directly.
  • url — your canonical homepage URL with https.
  • logo — an absolute URL to your logo image. Helps AI match your brand identity.
  • sameAs — links to your profiles on GitHub, LinkedIn, Twitter, Crunchbase, etc. This helps AI systems verify your identity and cross-reference information.

WebApplication example (for SaaS tools)

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "WebApplication",
  "name": "Builder Wrench",
  "url": "https://builderwrench.com",
  "description": "Developer tools for AI standards: llms.txt, MCP, and A2A.",
  "applicationCategory": "DeveloperApplication",
  "operatingSystem": "All",
  "offers": {
    "@type": "Offer",
    "price": "0",
    "priceCurrency": "USD"
  }
}
</script>

LocalBusiness example

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "LocalBusiness",
  "name": "Sunrise Bakery",
  "url": "https://sunrisebakery.com",
  "telephone": "+1-212-555-0147",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "456 Broadway",
    "addressLocality": "New York",
    "addressRegion": "NY",
    "postalCode": "10013",
    "addressCountry": "US"
  },
  "openingHours": "Mo-Fr 07:00-18:00"
}
</script>

Common mistakes

  • Vague descriptions: "We help businesses grow" tells AI nothing specific. Write what you actually do: "Project management software for engineering teams."
  • Wrong @type: using Organization for a product means AI misses fields like applicationCategory and offers.
  • Relative URLs: all URLs in structured data must be absolute (starting with https://).
  • Stale data: if your description or pricing changes, update the structured data. AI will cite it as-is.
  • Only on homepage: structured data on the homepage is the minimum. For content sites, add Article or FAQPage schema to individual pages.

Validating your structured data

Before publishing, test your JSON-LD at validator.schema.org or Google's Rich Results Test. Both will flag missing required fields and syntax errors.

How it fits into overall AI readiness

Structured data is worth 25 points in the AI Readiness scoring (out of 100). It works alongside llms.txt: llms.txt gives AI a prose summary of your site, structured data gives it machine-readable facts. Together they cover both how AI understands you and how it describes you to users.

llms.txt guide → | Full AI readiness checklist →

More Guides

View all guides →