Schema Markup for AI Agents
JSON-LD Patterns That Matter
Not all structured data is created equal. AI agents rely on specific JSON-LD types to understand who you are, what you offer, and how to interact with your product. This guide covers the schemas that actually move the needle.
Why Schema Matters for AI Agents
When a human visits your website, they interpret visual cues — logos, headlines, pricing cards, navigation menus. When an AI agent visits, it reads HTML. Structured data bridges this gap by embedding machine-readable metadata directly in your pages.
JSON-LD (JavaScript Object Notation for Linked Data) is the format Google recommends and AI agents prefer. It lives in a <script type="application/ld+json"> tag and does not affect your visual layout. Agents parse it to extract facts: your company name, product pricing, API endpoints, FAQ answers, and more.
Without structured data, agents must infer meaning from raw HTML — a process that is slow, error-prone, and token-expensive. With it, they get unambiguous, pre-structured facts in a format designed for machines.
Organization Schema
The Organization schema tells agents who you are. It is the most fundamental schema type and should appear on your homepage. Agents use it to identify your brand, find your social profiles, and understand your contact points.
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "Your Company",
"url": "https://example.com",
"logo": "https://example.com/logo.png",
"description": "One-sentence description of your company.",
"sameAs": [
"https://twitter.com/yourcompany",
"https://github.com/yourcompany",
"https://linkedin.com/company/yourcompany"
],
"contactPoint": {
"@type": "ContactPoint",
"contactType": "customer support",
"url": "https://example.com/support"
}
}Agent impact:When ChatGPT or Perplexity is asked "What does [Your Company] do?" — this schema provides the answer directly, reducing hallucination risk.
Product and WebApplication Schema
For SaaS companies, SoftwareApplication is more specific than generic Product schema. It lets you declare your application category, operating systems, pricing tiers, and feature list — all structured for agent consumption.
{
"@context": "https://schema.org",
"@type": "SoftwareApplication",
"name": "Your Product",
"applicationCategory": "DeveloperApplication",
"operatingSystem": ["Web Browser", "macOS", "Linux"],
"offers": [
{
"@type": "Offer",
"name": "Free Plan",
"price": "0",
"priceCurrency": "USD"
},
{
"@type": "Offer",
"name": "Pro Plan",
"price": "29",
"priceCurrency": "USD",
"billingIncrement": "month"
}
]
}Agent impact:When an agent is asked to "compare pricing for [category] tools," structured pricing data means it can include your product accurately instead of guessing or omitting you.
FAQPage Schema
FAQPage is disproportionately valuable for AI agents. When a user asks an agent a question that matches one of your FAQ entries, the agent can pull the answer directly from your structured data — often verbatim.
This is not theoretical. Perplexity and Google AI Overviews both extract FAQ schema answers as primary sources. If your competitors have FAQ schema and you do not, their answers appear and yours do not.
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "How much does Your Product cost?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Free for up to 1,000 requests/month. Pro plan is $29/month."
}
},
{
"@type": "Question",
"name": "Does Your Product offer an API?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Yes. RESTful API with OpenAPI spec at /openapi.json."
}
}
]
}WebAPI and Action Schemas
For sites with APIs, the WebAPI schema type tells agents that your site offers programmatic access. Combined with an OpenAPI spec link, it gives agents everything they need to start making API calls.
Action schemas (SearchAction, OrderAction, etc.) define specific things agents can do. A SearchAction on your homepage tells agents that your site has a search function and how to use it. An OrderAction on a product page tells agents that purchases are possible.
{
"@context": "https://schema.org",
"@type": "WebAPI",
"name": "Your Company API",
"url": "https://api.example.com",
"documentation": "https://example.com/docs/api",
"termsOfService": "https://example.com/terms",
"provider": {
"@type": "Organization",
"name": "Your Company"
}
}Common Schema Mistakes
Duplicate schemas
Multiple Organization schemas on the same page confuse agents. Use one authoritative schema per type per page.
Missing required fields
A Product schema without price or an Organization without name is worse than no schema — it signals low quality.
Stale pricing data
If your schema says "$19/month" but your page says "$29/month," agents will surface the wrong number. Keep schemas in sync with content.
Schema spam
Adding irrelevant schema types (e.g., LocalBusiness for a fully-remote SaaS) reduces trust signals. Only use types that accurately describe your business.
Client-side only rendering
If your JSON-LD is injected by JavaScript after page load, many agents will not see it. Ensure schemas render server-side.
No Article schema on content pages
Blog posts and guides without Article schema miss out on AI citation. Add headline, author, datePublished at minimum.
Schema and Your AX Score
Schema coverage is the primary signal in the Semantic Structure dimension (15% of AX score). The audit checks for valid JSON-LD on every crawled page, verifies required fields, detects duplicates, and validates against Schema.org specifications. Sites with comprehensive, accurate schema across their key pages score 12-15 points on this dimension alone.
Audit Your Schema Coverage
The AX Audit crawls your site, extracts every JSON-LD block, validates it against Schema.org, and tells you exactly what is missing, broken, or outdated.
Audit Your Schema CoverageFree audit. No signup required.