Schema markup

Schema markup is a way of saying, in a form a parser cannot misread, what a page is about. Prose says "Robots.txt is a file that tells crawlers what to fetch". Markup says: this is a DefinedTerm, its name is "Robots.txt", here is its definition, it belongs to this glossary. Search engines use it to understand pages and, for some types, to show richer results.

How does schema markup work?

The vocabulary is Schema.org, a shared set of types (Article, Product, Organization, FAQPage, DefinedTerm and hundreds more) and their properties. You describe the page in that vocabulary and embed the description in the HTML. Google supports three syntaxes and recommends one, JSON-LD: a single script block, separate from the visible markup, easy to generate from the same data that renders the page.

{
  "@context": "https://schema.org",
  "@type": "DefinedTerm",
  "name": "Schema markup",
  "url": "https://attensira.com/glossary/schema-markup",
  "termCode": "schema-markup",
  "inDefinedTermSet": {
    "@type": "DefinedTermSet",
    "name": "Attensira AI SEO & GEO Glossary",
    "url": "https://attensira.com/glossary"
  }
}

That is a cut-down version of what this page carries: here the DefinedTerm sits inside a WebPage block, one of five JSON-LD blocks the page emits (Organization, WebPage, BreadcrumbList, FAQPage and the author's Person).

What does schema markup actually do?

For classic search, two things. It helps the engine understand the page, and for documented types it makes the page eligible for a rich result: review stars, product prices, breadcrumbs, video thumbnails. Eligible, not entitled. Google says correct markup does not guarantee a rich result.

Eligibility also changes. Google's FAQ rich results were narrowed to well-known government and health sites, then stopped appearing in Google Search on 7 May 2026. We keep FAQPage on our pages because it states, in a form any parser reads, which questions the page answers. We do not claim it earns anything in Google any more.

Does schema markup help with AI answers?

Not as a requirement, on the documented evidence. Google's page on its AI features says there is no special schema.org markup needed to appear in AI Overviews or AI Mode; a page needs to be indexed and eligible for a snippet. OpenAI, Anthropic and Perplexity document crawler access, not markup.

The answers themselves still recommend it. ChatGPT's answer of 3 October 2026 to "AI search optimization for seed-stage startups", read from our own workspace, puts "Add schema markup" in a seed-stage startup's first 30 days and lists FAQ schema among the types, five months after Google stopped showing FAQ rich results. Both hold: markup helps a parser understand a page, and nothing documented makes it a ticket into an answer.

Our own product writes the same conclusion into its rules. When Attensira's crew adds JSON-LD to a customer's page, the change carries a fixed note: "hygiene, not a lever: we do not expect this alone to move an answer." Markup is worth having correct. It is not why an engine names you.

How do you know your markup is still there?

You usually find out late. Markup is invisible on the page, so a template change or a rendering fault can drop a block without anyone seeing a difference, and the first sign is a rich result that stops appearing.

The fix is a dated record. Attensira re-reads the pages in a workspace's inventory and keeps the JSON-LD @type values each version declares. When that set changes, say FAQPage disappears, the read is stored as a new version of the page, with its date. The page history then reports JSON-LD as present, absent or not captured, and it keeps the last two apart: a version read before metadata was recorded cannot say the markup was missing, only that nobody looked. With a recorded "before", a lost rich result or a dropped citation has a date to start from.

What do people get wrong about schema markup?

• Marking up what the page does not show. Google says not to mark up content hidden from readers. Breaking that rule can cost rich results through a manual action.
• Treating it as a ranking or citation trick. It describes; it does not persuade.
• Copying a template and leaving stale values. A price or rating in markup that no longer matches the page is a false claim in machine-readable form.
• Validating once. Templates change and markup breaks silently. Check it again after every template release, not once at launch.

How does Attensira handle structured data?

The free schema markup generator writes JSON-LD only from what a page actually says, and the schema validator checks existing blocks against each type's required properties. In a workspace, adding structured data is one of the few changes Attensira classes as safe. A price, rating or review count inside structured data is the exception: it is a claim, and like every claim it waits for your approval in every mode, because a price inside a schema block is still a price.