Databricks Makes Genie App for Slack Generally Available

Databricks on November 12, 2026 announced general availability of its Genie app for Slack, enabling employees to ask permission-aware data questions, analyze files, create charts, schedule analyses and approve actions in connected tools without leaving Slack.

Users can send a direct message to Genie or mention @Genie in a channel or group DM. Replies stay within the relevant conversation or thread, cite their sources and let users open the exchange in Databricks for more context. Connected enterprise knowledge sources can include Google Drive, Confluence and Jira.

Genie One is the default scope, routing requests across the connected data estate. A conversation or channel can instead be pinned to a curated Genie Agent so that answers remain within that agent’s data and instructions. Channel owners can select the connected Databricks workspace, public or private response visibility and the Genie scope.

From Answers to Approved Actions

Through MCP connectors to external tools, Genie can create or modify documents, comment on tickets and update records with the requesting user’s credentials. It prepares those changes for approval before execution. Only the person who asked the question can approve or deny a tool action; where a scope selector is offered, approval can apply once, always in that chat or always.

Recurring analyses can be created in natural language or through a scheduling button beneath an answer, with results delivered to the requester’s Genie DM. The app also accepts uploaded files, can return documents such as PDFs and renders tables and charts in Slack. The product documentation notes one presentation restriction: charts do not appear when response visibility is private.

Permissions and Visibility Are Separate Controls

Genie evaluates each request under the asker’s identity, with Unity Catalog and connected-source permissions limiting the information it can use. That retrieval control does not restrict the Slack audience by itself. Public responses, the default setting, are visible to all members of a channel or group DM, including members who cannot access the underlying data.

Visibility can be set for an individual, a channel or an entire workspace. Workspace settings take precedence over channel-wide settings, which in turn override personal choices; administrators can also require every response to be private. The result is a distinction administrators must manage explicitly: user permissions govern what Genie may retrieve, while collaboration settings govern who may read the answer.

Question processing and answer generation occur in the customer’s Databricks workspace, and conversation history is stored in that workspace’s region. The central service relaying messages with Slack does not persist or log request or answer content; it stores only the mapping between the Slack and Databricks workspaces. The workspace’s geography and cross-geography processing setting determine where large language model processing occurs.

Deployment Controls and Limitations

Unity Gateway provides centralized MCP policies, cost controls and monitoring, including traces of Genie’s actions and the users who requested them. Audit logs identify Slack as the request origin, separately from Databricks web and Microsoft Teams activity. Installation requires a Slack workspace administrator, after which users authenticate to Databricks and select a workspace. Workspaces using Private Link additionally require an account administrator to configure a context-based ingress rule for the app.

The app automatically selects a SQL warehouse that the user is authorized to access rather than offering a compute choice. Users cannot create Genie agents from Slack, and pinning a specific agent affects only new conversations. Databricks warns that its agents can generate inaccurate responses and tells users to verify output before acting on it.

The release turns Slack into another front end for Databricks’ existing governance model. Employees can query and act on business context inside their conversations, while administrators retain control over retrieval permissions, external actions and the audience for the resulting analysis.