Model vs. Harness: Where AI Agent Controls Belong
A comparison of what AI models should learn versus what agent infrastructure must enforce. See how reasoning, permissions, isolation, budgets, and observability split across the boundary.
DatabricksToussaint Webb, Sydney SundellAnnouncement
The Databricks Genie app for Slack is now generally available, allowing teams to ask data questions and receive governed answers directly in Slack channels, threads, or DMs. Genie uses the asker’s Unity Catalog and source permissions, so it only surfaces information that person is allowed to see. Every answer cites its sources, and users can click through to explore deeper in Genie One. Beyond answering questions, Genie can take action via MCP connectors, schedule recurring analyses, and work with uploaded files such as PDFs. Admins govern AI usage centrally in Unity Gateway, including MCP policies, cost controls, and monitoring, and Private Link workspaces are supported through context-based ingress policies.
Business questions often require data from scattered sources such as data platforms, documents, tickets, and meeting notes, causing delays while answers are pieced together. The Databricks Genie app for Slack aims to eliminate this friction by bringing governed AI answers directly into Slack channels, threads, and DMs.
The Databricks Genie app for Slack is installed from the Slack Marketplace and configured by an admin. Users then DM Genie or @mention it in a channel, sign in to Databricks, and pick a workspace. Genie answers in the thread using the asker’s Unity Catalog and source permissions, cites sources, and can take actions via MCP connectors, schedule recurring analyses, and work with uploaded files.
Genie in Slack answers questions using the identity and permissions of the person who asked, so it only surfaces data that user is allowed to see via Unity Catalog and connected sources.
Users can take action from Slack through MCP connectors, such as creating documents or updating tickets, with Genie preparing each change for approval before it runs.
Admins can govern Genie usage centrally in Unity Gateway, including MCP policies, cost controls, monitoring, and setting default response visibility for the whole workspace.
Yes. Genie One is the default, searching across the entire data estate. You can also pin a conversation or channel to a specific Genie Agent, which scopes answers to that agent's curated data and instructions.
Every question runs as the user who asked it, so Genie only uses data that person is allowed to see based on Unity Catalog and source permissions. Questions are processed inside your Databricks workspace, and the routing service does not store questions or answers.
Yes. A Databricks account admin must create a context-based ingress policy to allow secure connectivity between Slack and the Databricks workspace. See the documentation for setup steps.
Summaries are generated from the original by a language model. Check the original before relying on it.
See the idea
A comparison of what AI models should learn versus what agent infrastructure must enforce. See how reasoning, permissions, isolation, budgets, and observability split across the boundary.
Keep reading
The Good Stuff
Paid picks we use ourselves. Links carry a referral code — a subscription through one earns us a commission at no extra cost to you.
One useful email. Unsubscribe anytime.