What Schema.org is, and why search engines and AI assistants read it

Schema.org is a shared vocabulary that tells search engines and AI assistants what a page is about, so a machine does not have to guess.

The layer you do not see

A person looking at a web page sees a menu, a phone number, a photo of the building. A crawler sees HTML and has to work out what each piece is. Is "4.8" a price, a rating or a version number? Is the name at the top of the article its author or its subject?

Schema.org is a vocabulary for saying it outright. Google, Microsoft and Yahoo launched it in 2011, and Yandex joined later that year. It defines several hundred types, such as Organization, LocalBusiness, Article, Event, Product and FAQPage, and each type has its own properties. An Article has an author and a datePublished. A LocalBusiness has an address and opening hours. A Hotel has a checkinTime.

What Google does with it

Google reads Schema.org markup to decide whether a page can have a rich result, the search listing with extra detail such as an event date, a product price or a recipe's cooking time. Markup makes a page eligible. Google decides whether to show anything, and often it does not.

Google also retires rich results. The FAQ rich result, the list of expandable questions under a listing, went for almost all sites in 2023, and for the rest on 7 May 2026. FAQPage markup still says which text is a question and which is its answer, but it no longer changes how any site looks in Google.

The same markup helps Google work out who a business is, with its name, address, logo and social profiles. Google says Organization markup can influence things like the logo it shows and the knowledge panel, the box beside the results for a known business or person. It builds those panels from many sources, though, and no markup orders one.

What AI assistants do with it

This part is less settled, and I would be wary of anyone who claims certainty. ChatGPT search, Perplexity and Google's AI Overviews read pages and assemble answers from them. None of them publishes a rule saying structured data earns a citation.

The case for publishing it anyway is simple. JSON-LD states facts with no layout around them. If your opening hours sit in openingHoursSpecification, a program reading the page does not have to find them in a sidebar and guess the format. Less guessing leaves less room to get your details wrong. That is a reason to publish markup. It is not a promise that anyone will quote you.

The format to use

Schema.org can be written as Microdata, RDFa or JSON-LD. Google recommends JSON-LD, and I would too. It sits in its own <script type="application/ld+json"> block, apart from the visible HTML, so a template change cannot break it.

A small Organization block looks like this:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Acme Hotel Belgrade",
  "url": "https://acme-hotel.com",
  "telephone": "+381 11 123 4567",
  "address": {
    "@type": "PostalAddress",
    "addressCountry": "RS",
    "addressLocality": "Belgrade"
  }
}

Every property you fill in is one thing fewer for a machine to infer. Keep them true. Markup that disagrees with what the page shows breaks Google's structured data guidelines.

Doing it on Joomla

Writing JSON-LD by hand is not hard. Keeping it right is. A hotel's business block with hours, check-in times and amenities runs to dozens of lines, and every new phone number or change of hours means editing it again.

AI Boost for Joomla builds this markup on the server from settings you fill in once. On its Structured Data screen you pick one of 36 site types in the "Schema Type" field, and the screen then shows the fields that type can publish. A restaurant gets cuisine and price range, a hotel gets check-in and check-out times. Once you switch it on and fill in "Organisation Name", every indexable page carries one JSON-LD block with your business and the page itself, plus a breadcrumb trail and Article or BlogPosting markup where they apply.

It does not publish Product or AggregateRating markup, and it cannot make Google show a rich result. It makes each page state its facts, and a page that states them plainly is the most a plugin can give you.