Enforce consistent, secure tagging across data and AI assets with Governed Tags in Unity Catalog (Public Preview)

Today, we are excited to announce Governed Tags in Public Previewacross all regions on AWS, Azure, and Google Cloud!

Tags in Unity Catalog (UC)are widely used for discovery, cost tracking, and access management - but as adoption grows, inconsistent tagging can create problems. One team tags a dataset "finance," another uses "fin," and suddenly, searches, audits, and governance policies are incomplete. This makes compliance harder, weakens security enforcement, and complicates cost attribution.

Governed Tags solve this by introducing account-level tag policies where account admins and data stewards can define who can use a tag, what values are allowed, and where the tag can be applied. These tags can be applied to UC objects such as tables, views, volumes, catalogs, schemas, and columns, with support for workspace-level objects like dashboards and notebooks coming soon. They work across every workspace in an account, ensuring consistent metadata for discovery, security, and cost reporting at scale.

In this blog, we cover core use cases, how they work with existing tags, and what's next.

When to use Governed Tags

Tags are most valuable when they solve everyday problems for teams. Governed Tags make tagging more reliable and consistent, which opens up practical benefits across discovery, governance, and cost management. Here are three common ways organizations can use them:

Discovery

Consistent tagging makes data easier to find. With Governed Tags, a user searching for "marketing" will see every marketing dataset across the metastore, without missing assets labeled "mktg." Teams can still use free-form tags for ad hoc organization, like marking a dataset "ML_ready." In the future, Governed Tags will also support certifications and deprecations so users can quickly identify trusted or outdated data.

Governance and compliance

Governed Tags simplify classification and audits for scalable governance across your entire Databricks account and workspaces. With Data Classificationenabled, sensitive columns are automatically tagged, scaling discovery and compliance of sensitive data. Combined with Attribute-Based Access Control (ABAC),policies, this ensures that sensitive data (such as PII) is only accessible to the right users across your data platform.

Cost attribution and optimization

Tags can also track usage and costs. For example, tagging datasets with "costcenter: finance" makes it easy to attribute storage and query spend to the finance team. This visibility supports accurate chargebacks and better budgeting. In upcoming releases, Governed Tags will integrate with serverless budget policiesto further streamline cost control.