Structured data for AI search, explained simply
In short
Structured data is a block of hidden code that states your key facts, like name, address, prices and authors, in a format machines read without guessing. It helps search engines understand a page, but Google says no special structured data is needed for its AI features, so treat it as good housekeeping.
Structured data is a small block of code in your page that lists your key facts in a standard format, so machines do not have to guess. It says "this is a bike shop, this is its address, it opens at 9:00" in a way any program can read.
It helps search engines understand your pages. It does not get you into AI answers on its own. Google says plainly that "there's also no special schema.org structured data that you need to add" to appear in its AI features. So think of it as tidy labelling.
What is structured data, in plain words?
Google describes it as "a standardized format for providing information about a page and classifying the page content."
The vocabulary comes from schema.org, a shared dictionary of types like LocalBusiness, Product and Article, each with named properties. The usual way to add it is JSON-LD: a short block of text inside a <script type="application/ld+json"> tag. Visitors never see it.
Google supports three formats and recommends JSON-LD because it is the easiest to add and maintain.
What does structured data do for AI answers, and what does it not do?
| It does | It does not |
|---|---|
| State your facts without ambiguity: which number is the phone, which date is the publish date | Guarantee a citation in ChatGPT, Gemini or any other assistant |
| Help search engines understand a page, and can make some rich results possible | Replace clear visible text; Google wants the two to match |
| Tie a page to a named business or author | Make up for blocked crawlers or text hidden in scripts |
Give Apple a way to flag paywalled content: its isAccessibleForFree property affects use in AI answers | Work if it is broken or describes things visitors cannot see |
Our view: add it because it is cheap and removes doubt about basic facts. Do not expect it to move AI citations by itself, because no AI company we checked has said it does.
What should a local business add?
A LocalBusiness block, or a more specific type such as BikeStore, Dentist or Plumber. Google lists only name and address as required. Phone, opening hours, web address and price range are recommended.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "BikeStore",
"name": "Harbour Bikes",
"url": "https://www.example.com/",
"telephone": "+44 117 496 0000",
"priceRange": "££",
"address": {
"@type": "PostalAddress",
"streetAddress": "12 Quay Street",
"addressLocality": "Bristol",
"postalCode": "BS1 4XX",
"addressCountry": "GB"
},
"openingHoursSpecification": [
{
"@type": "OpeningHoursSpecification",
"dayOfWeek": ["Tuesday", "Wednesday", "Thursday", "Friday", "Saturday"],
"opens": "09:00",
"closes": "17:30"
}
]
}
</script>
Swap in your own details. The phone number and postcode above are placeholders. Use the same name, address and phone as on your page and your business profiles.
What about products in an online shop?
Use Product with an Offer. Google's product snippet rules need a name and at least one of offers, review or aggregateRating. Price currency and availability are recommended.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Product",
"name": "Commuter hybrid bike, 54 cm frame",
"description": "Aluminium hybrid bike with 8 gears, mudguards and a rear rack.",
"image": "https://www.example.com/images/commuter-hybrid.jpg",
"sku": "HB-COM-54",
"brand": { "@type": "Brand", "name": "Harbour Bikes" },
"offers": {
"@type": "Offer",
"url": "https://www.example.com/bikes/commuter-hybrid",
"price": "649.00",
"priceCurrency": "GBP",
"availability": "https://schema.org/InStock"
}
}
</script>
Only add review or rating data if the reviews are real and shown on the page.
How should articles and guides be marked up?
Use Article, BlogPosting or NewsArticle. Google lists no required properties but recommends a headline, image, dates and author. For the author, use a Person with a name and a link to a page about them.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "How often should you service a road bike?",
"image": "https://www.example.com/images/road-bike-service.jpg",
"datePublished": "2026-11-16T09:00:00+00:00",
"dateModified": "2026-11-16T09:00:00+00:00",
"author": {
"@type": "Person",
"name": "Sam Example",
"url": "https://www.example.com/about/sam"
},
"publisher": {
"@type": "Organization",
"name": "Harbour Bikes",
"url": "https://www.example.com/"
}
}
</script>
Dates matter. Showing "published" and "last updated" on the page, and in the markup, tells readers and machines how fresh the advice is.
Is FAQ markup still worth adding?
Less than it was. Google stopped showing FAQ rich results in May 2026. Its changelog says the feature "will no longer appear in Google Search results starting May 7, 2026."
FAQPage is still a valid schema.org type, and the visible questions and answers are still useful to readers and to AI tools. If your FAQ section exists anyway, the markup is a few minutes' work:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "How long does a standard bike service take?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Most standard services are ready the same day if booked before 11:00."
}
}
]
}
</script>
Our view: write the FAQ for people first. Add the markup only if it costs you little.
How do I check my structured data?
- Paste your page address into Google's Rich Results Test or the Schema Markup Validator.
- Fix any errors it lists.
- Compare each value with the visible page. A different phone number or a price that is not on the page is a problem.
Our free scanner checks whether your homepage has a business description in structured data, whether that markup has errors, and whether pages say what type of page they are. It reports what is present and valid; it cannot tell you how any engine weighs it. The methodology page explains each check. For the wider checklist, see how to check your site in 15 minutes.
Frequently asked questions
Do I need a developer to add structured data?
Not always. Many website builders and shop platforms add basic markup for you, and some have a settings page for business details. For custom sites, it is usually a small job for a developer.
Can structured data hurt my site?
Markup that is misleading, or describes content visitors cannot see, can make a page ineligible for rich results, according to Google. Accurate markup that matches the page carries no such risk.
Which type should I use if my business does not fit?
Pick the most specific type that is true, and fall back to LocalBusiness or Organization. Schema.org lists every type and its subtypes.
Is JSON-LD the same as llms.txt?
No. JSON-LD sits inside each page and uses the schema.org vocabulary. llms.txt is a separate file at the root of your site. See what llms.txt is.
Sources
- Google Search Central: Introduction to structured data
- Google Search Central: AI features and your website
- Google Search Central: Local business structured data
- Google Search Central: Product snippet structured data
- Google Search Central: Article structured data
- Google Search Central: Latest updates (FAQ rich result deprecation, 2026)
- Schema.org: LocalBusiness
- Schema.org: Product
- Schema.org: Article
- Schema.org: FAQPage
- Google: Rich Results Test
- Schema.org: Schema Markup Validator
- Apple Support: About Applebot