Does schema markup help with AI search?

What schema actually does

Schema markup is structured code, usually JSON-LD, that labels the facts on a page. Instead of a search engine guessing that “403-555-0100” is a phone number, the markup says so.

Google uses it to understand pages, connect them to its Knowledge Graph and show rich results like review stars or event dates. Because Google’s AI features sit on top of its search systems, anything that helps Google understand you can help there too.

Where the evidence is thinner

For other AI assistants, it’s less clear. Many fetch a page and read the text, and there’s limited public detail on how much weight they give to structured data. Google has said no special schema is needed to appear in AI Overviews.

So I’d be wary of anyone claiming schema is the key to AI search. The visible content does most of the work. Schema reinforces it, and it should never say anything the page doesn’t.

Which types are worth adding

• Organization or LocalBusiness: name, logo, address, phone, hours, service area and links to your official profiles.
• Service or Product: what you offer, where relevant.
• BreadcrumbList: how pages relate to each other.
• Article: author and date on blog posts.
• Review data only where it follows Google’s rules. Self-serving review stars on your own business usually aren’t eligible.

A schema generator gets you a valid starting point. Test it with Google’s Rich Results Test.

I add Organization or LocalBusiness schema sitewide, plus breadcrumbs and page-appropriate types, generated from the same data as the visible content so they can’t disagree. The sameAs links to the business’s official profiles help tie everything to one entity.

I don’t sell schema as an AI search shortcut. It’s a small, cheap piece of being clear about who you are. The content and reputation are what get you named.