Hello @MauriceDekker ,
Good question, and you've framed it well — UC has genuinely gotten strong for physical data assets and metrics, so reports and business terminology feel like the obvious next frontier. Here's how I see customers approaching this today, along with where the real limits are.
On governing Power BI reports in UC:
Honest answer: You can't register or permission a report directly inside UC today. The pattern that tends to work is keeping the semantic layer centralized in Databricks — curated gold tables and metric views with clear names, descriptions, and owners — and requiring Power BI reports to connect live to those UC-governed objects rather than letting each team build ad hoc datasets on the side.
Access control ends up working in two layers: UC permissions on the underlying tables/views, and Power BI workspace/report permissions on top. If someone doesn't have access to a dataset in UC, they can't see that data in any report consuming it. That's your real governance enforcement point.
For the report catalog piece, since reports aren't native UC assets, many teams build a lightweight registry as a Delta table — something like governance.report_registry with columns for report name, workspace, owner, business domain, primary datasets, PII classification, and status. A small scheduled job pulling from the Power BI REST API keeps it current. Analysts can then browse certified reports and see exactly which governed datasets they reference. Not glamorous, but it works.
On a business glossary in UC:
No built-in glossary UI exists today — no dedicated object type, no lifecycle workflows out of the box. What works is building it from native primitives. A governance schema (e.g., governance.business_glossary) with a terms table covering name, definition, business owner, domain, status, and synonyms gets you most of the way there. Add a relationships table if you need broader/narrower/related term linkage.
From there, UC column and table comments handle concise business descriptions, and tags link assets back to glossary terms — tagging a column with business_term = "Net Revenue" or data_classification = "Confidential", for example. You can then query UC system tables or information_schema to find where a term is actually implemented and which columns carry a given classification. Restrict writes to a small stewardship group, open reads broadly, and optionally surface it through a Databricks SQL dashboard so business users don't have to query tables directly.
If you're already running Purview, Collibra, or Alation, a reasonable model is keeping that as your primary glossary UI and syncing terms and classifications into UC via their connectors. UC enforces access; the external tool handles the governance workflow and UX.
Hope this helps.
Cheers, Louis
Want anything tweaked before you post?
