JSON-LD structured data for AI crawlers

Structured data (JSON-LD) is a small block of machine-readable facts embedded in your pages. Search engines have used it for a decade to power rich results and knowledge panels. For AI visibility it plays a quieter but real role: it's the cleanest signal you can give a machine about what you are (your name, category, pricing and relationships) without hoping a parser gets it right from your marketing copy.

Honest framing: nobody publishes the weights an assistant gives to JSON-LD. What we know is that answer engines are built on web crawlers and search indexes that demonstrably do parse it, that it costs minutes to add, and that ambiguous entities lose to unambiguous ones. It's hygiene: do it once, do it truthfully.

The three blocks most brands should ship

1. Organization: who you are. One per site, on every page (layout-level):

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "@id": "https://example.com/#org",
  "name": "Example",
  "url": "https://example.com",
  "description": "One factual sentence: category + who it's for.",
  "sameAs": [
    "https://www.linkedin.com/company/example",
    "https://github.com/example"
  ]
}

sameAs links are underrated: they let machines connect your site to your profiles on platforms they already trust.

2. Product / SoftwareApplication: what you sell. On your homepage or product page, including real pricing:

{
  "@context": "https://schema.org",
  "@type": "SoftwareApplication",
  "name": "Example",
  "applicationCategory": "BusinessApplication",
  "operatingSystem": "Web",
  "description": "What it does, stated plainly.",
  "offers": {
    "@type": "Offer",
    "price": "39",
    "priceCurrency": "USD"
  },
  "publisher": { "@id": "https://example.com/#org" }
}

Physical products use Product with offers, brand and aggregateRating (only with real review data behind it).

3. FAQPage: the questions you answer. On pricing and comparison pages, mirror the questions buyers ask assistants:

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [{
    "@type": "Question",
    "name": "Who is Example for?",
    "acceptedAnswer": {
      "@type": "Answer",
      "text": "Example is built for small agencies managing multiple client projects."
    }
  }]
}

Rules that keep it working

• Match the visible page. Structured data that contradicts on-page content is the one pattern that gets penalized rather than ignored.
• Be truthful. No invented ratings, no fake review counts. The point is to be the reliable source of facts about yourself.
• Use one `@graph` when combining several blocks on a page, with @id references tying them together, rather than scattering disconnected scripts.
• Validate with Google's Rich Results Test or schema.org's validator after every change: one trailing comma silently voids the block.

Structured data tells machines what you are; it can't make them recommend you. For that, see how citations actually work, and check what assistants say about you today with the free scan.