What it means
Metadata is data that describes other data, such as a document's title, author, date, version, language, status, access level or subject. It helps people and systems find, filter, interpret and manage information.
Real-world example
An AI assistant for a regional government answers a question about travel expense rules using a 2019 guideline, although a 2024 version exists. Both documents had the same title, and neither had a reliable "valid from" date or "superseded" status. Adding consistent version and validity metadata lets the system prefer current documents and cite their dates.
Why it matters for AI adoption
Good metadata is one of the most effective ways to improve AI search and answer quality. It also enables access control, retention rules and audit trails. Many organisations underestimate how much metadata clean-up AI projects require.
Questions to ask
- Do our documents have reliable metadata for version, validity, owner and access level?
- Does the AI system use metadata when retrieving and ranking information?