Key points
- Structured data is machine-readable information, usually written in JSON-LD using schema.org vocabulary, that describes what a page is, who wrote it and who published it.
- Google says there is no special structured data needed for AI Overviews or AI Mode, but that existing markup should be accurate and match the visible page.
- Its real value for AI visibility is identity: linking articles to real authors and to a publisher whose name, founding date and profiles are consistent everywhere.
Few things in AI visibility are oversold as often as schema markup. Adding code to a page will not make an AI assistant cite it. But accurate structured data does help machines answer basic questions about a page with confidence, and inconsistent data can do harm. This article is part of our AI search visibility guide.
What structured data is
Structured data is information about a page written in a standard, machine-readable format. On most sites it is JSON-LD code in the page’s HTML, using the shared vocabulary maintained at (schema.org). It can say, for example, that a page is a news article, that it was written by a named person, published by a named organisation on a given date, and that it is about a particular subject.
What Google says
Google’s guidance on its AI features is explicit: “There’s also no special schema.org structured data that you need to add.” It asks site owners to keep structured data accurate and consistent with the visible content of the page (Google Search Central). Markup that describes things readers cannot see on the page is, at best, ignored.
What actually helps
- Article markup with a real author. Mark up the headline, dates, author and publisher, and link the author to a profile page that shows who they are.
- A consistent organisation. Describe your publisher once, with the same name, alternate names, founding date, logo and official profiles on every page. Link it to your Wikidata entry if you have one using
sameAs. - Stable identifiers. Give the organisation, authors and recurring series a fixed
@idso every page refers to the same entity rather than describing it afresh. - About and mentions. Link the main subjects of an article to well-known reference pages, so there is no doubt which “Mercury” or “Apple” you mean.
- FAQ markup only for visible FAQs. If a page has a real question-and-answer section, marking it up makes the answers easy to extract.
Common mistakes
- Conflicting descriptions. Several plugins each describing the same organisation with different details. Machines then see contradictions.
- Markup that does not match the page. Ratings, prices or FAQs in the code that readers cannot see.
- Duplicate article markup. Two plugins each adding their own Article schema to the same page.
- Treating schema as a ranking trick. It describes content; it does not substitute for it.
A worked example: fixing our own conflicts
When we audited Newsinterpretation in October 2026, we found three descriptions of the same organisation on every page, added by three different tools, with three different founding dates and places. We corrected them so that all three now match our Wikidata entry (Q141800919): the same names, an April 2019 launch date and the same official profiles. NI Geovanta, the plugin we use, holds the main organisation description; our SEO plugin handles article markup. One owner for each type of data is the simplest way to avoid contradictions.
For how this fits with crawl access and content quality, see our article on generative engine optimisation.
Frequently asked questions
Does schema markup help with AI search?
Indirectly. Google says no special structured data is needed for its AI features, but accurate markup helps AI systems identify a page’s author, publisher and subject, and consistent organisation data reduces confusion about who you are.
Which schema types matter most for publishers?
Article or NewsArticle with author and publisher, Person for authors, Organization or NewsMediaOrganization for the publisher, and FAQPage for visible question-and-answer sections.
What is sameAs in structured data?
sameAs links an entity in your markup to the same entity elsewhere, such as a Wikidata entry or official social media profile, so machines can confirm they refer to the same thing.
Can wrong structured data hurt?
Yes. Markup that conflicts with the page or with other markup on the site creates contradictions, and Google asks for markup to match visible content.
Last reviewed: 22 October 2026. Sources are linked inline.
Disclosure: NI Geovanta is developed by Newsinterpretation, the publisher of this article. It is used here as a worked example; other tools can perform similar checks.
