Structured data expresses page information in a defined format. Websites often use JSON-LD and the schema.org vocabulary. Its use depends on the consuming product and supported feature. Keep every stated fact consistent with the visible page.Google's introduction to structured data and schema.org's getting-started guide cover the format and the vocabulary.
Identifying the subject
Organization markup can identify the business and link to profiles that describe the same organization. Add only accurate identifiers and relevant sameAs links. This documents identity; it does not guarantee that a model will resolve an ambiguous name correctly.
Keep important facts readable in the page itself. Where markup and feeds are supported, generate them from the same maintained information. Check for differences in price, availability, name, and URL.
Choosing supported types
Choose types that fit the page and a documented use: Organization for business information, Product and Offer for products, or Article for an article. Check each platform’s feature requirements. Do not add unsupported or irrelevant markup solely because a type exists.
What each platform actually reads
Google documents structured data for supported search features and says no special schema is required for AI Overviews or AI Mode. For another product, check its own documentation. A claim that all assistants consume every schema.org field requires evidence.The mythbusting section of Google's AI features guide, under "overfocusing on structured data".
Product feeds
Where a shopping platform accepts product feeds, keep identifiers, prices, availability, and destination URLs accurate. Feed eligibility and coverage are platform-specific. Products, Comparisons, and Buying Tasks covers the commercial work.The local business and ecommerce section of Google's AI features guide.
The llms.txt proposal
llms.txt is a proposal for presenting selected site information in a text file. Its existence does not establish that a given search product reads it or uses it in ranking.The llms.txt proposal site, which specifies the format and collects tooling.Google's AI features guide, mythbusting section: no need to create machine-readable files or Markdown to appear in Google Search, which ignores them.
Use it when a documented integration needs it or as a clearly labeled experiment. Prioritize accessible pages, accurate information, and supported data formats before an unverified visibility claim.
Validate facts as well as syntax
Choose one product or business page. Compare the visible name, identifier, price, availability, and URL with its structured data. Then check the same facts in any submitted feed. A validator can accept syntactically valid markup that still contains an old price.
For Google product features, follow the requirements for the intended result type. Product snippets and merchant listings have different uses. Supported markup makes a page eligible for a feature; it does not guarantee display.Google product structured data. Distinguishes product snippets and merchant listings and links their requirements.
Record three results separately: the markup parsed, the facts matched the page, and the relevant provider accepted the submitted data. None of those checks establishes that an assistant will recommend the product.