About this topic
Data governance fails in two directions. Either the tenant switches are all-or-nothing and the business routes around them, or nobody owns the definitions and every department reports a different number. On Microsoft Fabric and Power BI the controls now exist to do better, but they are spread across Purview, workspace roles, capacity settings, domains, policies and the semantic model itself, and they only work when someone has decided what the rules are.
This hub collects our governance and security writing. It begins with what data governance is and a Microsoft governance framework you can actually implement, then goes into the Fabric specifics: the security model and where each permission is set, rule-based Fabric policies, domains and data mesh, and Purview for lineage, classification and sensitivity labels. Power BI controls are covered in detail, including row-level security and workspace governance that still works at a hundred reports. Compliance articles explain what GDPR and POPIA actually require of a data platform. The newest section is AI governance: keeping Copilot away from stale or sensitive content, Purview DSPM for AI, and Entra Agent ID as the identity perimeter for agents.
Our Microsoft data governance service page describes how we run this work as a programme rather than a one-off audit.