Databricks has introduced new capabilities in Agent Bricks, the governed enterprise agent platform that runs AI agents directly on your business data with the right permissions and controls. It unifies data, models, and governance in one place so teams can move from AI pilots to production agents that understand your customers, operations, and policies.
“William Acosta, Head of Agentic AI Engineering at EchoStar, puts it: ‘With Agent Bricks, we’re not building one-off AI projects; we’re building an enterprise AI fabric.’”
What’s new
- Document Intelligence GA: Lets agent read PDFs and other documents and turn them into usable data.
- Custom Agents on Databricks Apps GA: Build and launch your own agents on serverless Databricks Apps.
- Agent Mode in Genie: Ask Genie deeper, multi-step questions about your business data and get rich answers.
- Knowledge Assistant GA: Create chatbots that answer questions from your internal documents while respecting metadata and permissions.
- AI Gateway + MCP-connected tools: Control and monitor how agents use models and external tools from one place.
- End-to-end evaluation in MLflow: Test how well your agents work and stay safe before and after going live.
In the full post, you’ll see how Agent Bricks brings multi-AI routing, identity-first governance, and rich business context together into a single platform so you can build, secure, and scale enterprise agents that operate on your data, not just examples.
Read the Full Post here 👈
