Structured data for AI is machine-readable information embedded in a page that unambiguously specifies what individual pieces of information mean – organisation name, address, price, availability, author, date of update, rating. For language models and search engines, it reduces uncertainty when interpreting content and makes it easier to link the company to an existing entity. It is worth marking up, in priority order, the organisation including its identification details and links to profiles, products with price and availability, services, content authors and events. However, the markup must not claim anything different from the visible content of the page, otherwise it is ignored or leads to a penalty. It does not replace the quality of the text – it is an aid for confirming facts that are actually on the page. After deployment, it is worth checking the markup with a testing tool and repeating the check after every major change to the website's template.
See also: Structured data validation, Building a company entity, Knowledge graph and brand knowledge panel.
