Which structured data do you need, and what can it not do?

Structured data is a standard format for describing a page and classifying its content. It tells search engines whether a page is a clinic, a product or an article, who runs it, and where.

Google accepts it in three formats, JSON-LD, Microdata and RDFa, and recommends JSON-LD where your setup allows. JSON-LD is a separate block of code that does not mix with the page's text, and most of what Google reads in it uses the schema.org vocabulary.

But Google says its own documentation, not schema.org, is the reference for how its Search behaves. schema.org has properties Google does not need, which other engines and platforms may use. The short definition is in the glossary.

This guide is part of the steps to lead AI answers, which covers what to prepare on your site before and after structured data.

It does not guarantee a citation. Google says it is not required for AI Overviews or AI Mode, with no special type to add, but it makes your business's identity clearer to machines.

Google's guide to optimising for its generative AI features says structured data is not required for generative AI search, while still recommending it so your page can be eligible for rich results, such as breadcrumbs or a product's price and rating.

Microsoft, in its advertising blog, describes schema as code that helps search engines and AI systems understand content, turning plain text into data machines can interpret with confidence. That is Microsoft's own statement about Copilot and Bing, and we have no measurement that confirms or contradicts it.

In practice, structured data confirms what the text says and does not replace it. If the answer to the question is not written on the page, the code will not write it for you. What brings a page closer to being cited is in what to change on your page so it cites you.

Choose the most specific type for your business. For a local business, Google recommends the most specific LocalBusiness subtype, not the generic Organization. Here are the types per sector, with their key fields.

We checked every type below on schema.org, and the ones with a rich result in Google's documentation. Sector first, then type, then fields:

• Clinic: MedicalClinic, or Dentist for a dental clinic, or Physician for a single doctor's practice. Fields: address, openingHoursSpecification and telephone, plus medicalSpecialty on MedicalClinic.
• Online store: Product with an Offer. Fields: price as a number, priceCurrency set to your currency's three-letter code, shippingDetails for delivery and hasMerchantReturnPolicy for returns.
• Hotel: Hotel. Fields: checkinTime, checkoutTime, amenityFeature for facilities and starRating for the official classification.
• Restaurant: Restaurant. Fields: servesCuisine, hasMenu, acceptsReservations and openingHoursSpecification.
• School: School. Fields: name, address and telephone; stages and fees stay in the page text.
• Law firm: LegalService. Fields: areaServed for the cities it serves, address and telephone.
• Developer or estate agent: RealEstateAgent. Fields: areaServed for districts and cities, address and telephone.
• Business software: SoftwareApplication. Fields: applicationCategory set to BusinessApplication, and offers with the price.

Five mistakes recur in Arabic sites' structured data, all easy to fix. The most serious is text in the code that visitors cannot find on the page, which alone breaks Google's guidelines.

• Text the page does not show: a price, rating or question in the code that visitors cannot find on the page. Google's guidelines are explicit: do not mark up content that is not visible to readers of the page.
• No language set: an Arabic page with no inLanguage of ar, or an English version carrying the same Arabic description. Give each version its own language and its own description.
• One spelling of the name: customers write the business name in Arabic one day and in Latin letters the next, while the code holds one spelling. Add the other in alternateName, one of the properties Google suggests for Organization.
• Unencoded Arabic URLs: a URL with Arabic letters belongs in the code percent-encoded, as the browser sends it, so an Arabic letter starts with %D8 or %D9 rather than appearing as itself.
• Ambiguous numbers and currency: a price in Eastern Arabic numerals, or with the currency's abbreviation inside the price field. The field expects a number in Western digits, with the currency in priceCurrency as its three-letter code.

Our own site is an example you can inspect now. Open any guide in the Rich Results Test or the Schema Markup Validator to find the types below, linked by @id, and every value in them that describes the guide itself is written on the page.

These are the types as we built them:

• Organization named تصدَّر, with its Latin spelling Tasadr, its unvowelled Arabic spelling and its domain tasadr.ai in alternateName, and its one-sentence definition in disambiguatingDescription, so no engine confuses it with the everyday Arabic verb.
• Article for every guide, written by the Tasadr (تصدَّر) team as a team within the same organisation, published by the Organization by its @id, with the published and updated dates you read under the title, and the sources listed at the end in its citation property.
• DefinedTermSet on the glossary, each term a DefinedTerm linked to the set.
• Report on the sample report: a real measurement report with names hidden, dated, about the engines that were asked.

With three free tools: Google's Rich Results Test to see which rich results a page qualifies for, the Schema Markup Validator for the code's validity, and Search Console's reports to follow it after publishing.

• Rich Results Test at search.google.com/test/rich-results: enter a published page's URL or paste its code, and it tells you which rich results the page can generate and which field is missing.
• Schema Markup Validator at validator.schema.org: schema.org's own tool, which checks the code and its vocabulary, including types with no rich result in Google, such as School.
• Search Console after publishing: the rich result reports count valid pages and pages with errors, and the URL Inspection tool confirms Google found the markup.

It guarantees neither a rich result nor a citation. Google does not guarantee your structured data will be shown even when it passes the test, and does not require it for generative AI features.

• It no longer earns FAQ rich results: Google announced that the FAQ rich result would not appear in Search from 7 May 2026, then removed its documentation on 15 June 2026.
• It does not replace the text: if the answer is not written on the page, the code cannot add it, and trying to breaks Google's guidelines.
• Getting it wrong costs you the rich result: Google says a structured data manual action removes a page's eligibility for rich results, without affecting how it ranks in Search.
• It does not correct your Google profile: your business facts on Google Business Profile are edited in the profile itself, as set out in your Google Business Profile.