Governing AI Agents with Unity Catalog and Unity Gateway on Databricks | Unity Catalog Demo
Description
Unity Catalog filters data before it reaches the agent, so two users asking the same question get different, governed answers. Watch Lars George demo it with two users, Alice in support and Bob in finance, using Unity Gateway policies (built-in and custom UDFs), tag-based row filters, and per-user retrieval over a customer table and knowledge docs. 🎯 What You'll Learn: • How Unity Gateway applies built-in policies for jailbreak, unsafe content, and sensitive data • How to add custom guardrails with Unity Catalog UDFs, like blocking bulk customer exports • How tag-based row filters control what each user group can see • How the same agent returns different answers per user based on their access ⏱️ CHAPTERS: 00:00 - Introduction: Alice, Bob, and governed results 00:50 - Gateway policies and custom UDF guardrails 02:00 - Row filters, tags, and user groups 04:15 - Same question, different answers per user 📚 Resources: • Get Started with Data Governance training (free): https://www.databricks.com/training/catalog/get-started-with-data-governance-on-databricks-4678 • Unity Catalog: https://www.databricks.com/product/unity-catalog • Unity Gateway: https://www.databricks.com/…
Description from YouTube. Full content on the video page.
More from Databricks
TutorialsDatabricks Pipelines: Write to One Unity Catalog Table from Multiple Sources with FLOWs
NewsHow Repsol Democratizes AI With Databricks Genie | Data + AI Executive Series: Fast 5
NewsBuilding Reliable Data Pipelines with Lakeflow on Databricks | Lakeflow Demo
Databricks Lakeflow provides a data intelligence platform for building and managing reliable data pipelines. The demonstration covers creating delta tables, ingesting data from Salesforce using Lakeflow Connect, and using Lakeflow Designer for data transformation and orchestration.