cd /news/ai-agents/genie-in-slack-is-now-generally-avai… · home › topics › ai-agents › article
[ARTICLE · art-169341] src=databricks.com ↗ pub= topic=ai-agents verified=true sentiment=↑ positive

Genie in Slack is now Generally Available: Governed AI cowork, right where work happens

Databricks announced the general availability of the Genie app for Slack, letting users DM Genie or @mention it in a channel to get answers grounded in governed business context from Unity Catalog and external tool source permissions. Genie runs every question with the asker's permissions, cites its sources, and can update documents and tickets through MCP connectors after user approval, as well as work with files and run analyses on a schedule. Each channel connects to a chosen Databricks workspace and can route questions to Genie One by default or to a pinned Genie Agent for domain-specific scope.

by read5 min views1 publishedNov 12, 2026
Genie in Slack is now Generally Available: Governed AI cowork, right where work happens
Image: Databricks Blog

Bring Genie into Slack to get governed answers, charts, and actions right in your flow of work.

by Toussaint Webb and Sydney Sundell • Genie in Slack is generally available: DM Genie or @mention it in a channel to get governed answers grounded in your business context without leaving Slack.

• From answers to action: Genie updates documents and tickets through MCP connectors after you approve, and works with files and runs analyses on a schedule.

• Governed by design: Every question runs with the asker's permissions, so Unity Catalog and source permissions apply in DMs and channels alike.

Many business questions can’t be started and answered within the same conversation. A manager asks her team in Slack why revenue growth slowed last month across active customers. To answer, someone needs to find the right dashboard, check it against the number finance reported, and search last week's meeting notes for the explanation. Two hours later, they paste their findings back into the thread, but the conversation has already moved on.

The delay doesn't come from the question being hard, it comes from having to piece together data from multiple scattered sources. The revenue numbers live in your data platform, the explanation for the slowdown lives in documents and tickets, the definition of “active customer” varies across teams, and decisions happen in Slack.

Today, we’re making it easier than ever for teams to answer questions right where they start, with the General Availability of the Databricks Genie app for Slack. Genie now answers in the channel, thread, or DM where questions are asked, with the same governed context and permissions it uses directly in Databricks.

The Databricks Genie app for Slack brings the power of Genie directly into your Slack workspace. You can start a DM with Genie or @mention it in a channel, and Genie responds directly in the thread, with answers grounded in the governed business context you have access to through Unity Catalog and any outside tool source permissions.

Every answer cites its sources, so you can scan the context Genie used to compile your results without leaving Slack, or click directly on a source to explore deeper in Genie One.

Teams can work with Genie together in any channel Genie has been added to. Anyone who has connected their Databricks account can type @Genie, and Genie replies directly in the thread with a response. Genie will only read the context of the channels it has been invited to, and responds to each question within the context of the channel it was asked in.

Genie answers in the thread it's called from, so follow-up questions keep their context and move the conversation along. Responses are visible to everyone in the channel by default, giving the team shared context for decisions. For more privacy, admins can set responses to be private to the person who asked as the default for the whole workspace.

Each channel has its own configuration. You choose which Databricks workspace the channel connects to, and whether questions go to Genie One or to a specific Genie Agent. By default, channels use Genie One, which routes questions across Genie Search, Genie Agents, and more, for a comprehensive view of results across your data estate. For domain-specific channels such as "Product FAQs" or "Employee Benefits,” you can pin a curated Genie Agent to a channel, so answers stay scoped to that agent's data and instructions.

Users can work privately with Genie right from a DM. You can ask questions about your data, search the enterprise knowledge sources you've connected, see results as tables and charts, upload files for Genie to analyze and take action in other tools from the same chat. DM answers use your permissions and stay private to you, just as a chat would in Genie One or Genie Agents.

Of course, a trusted answer is useful only if someone acts on it. With Genie embedded directly in Slack, users can turn insights into work right from their Slack workspace:

These same capabilities empower AI cowork directly in Genie One, so you can orchestrate and review your workflows from whichever system you prefer.

Genie in Slack follows the same governance as every other Genie surface. Every question runs using the identity of the person who asked it, across Unity Catalog and each connected source, so Genie uses only information that person is allowed to see. In channels, answers use the permissions of whoever asked, and each channel can choose whether responses are shared with the channel or sent privately to the asker. Workspace admins can set this default for the whole workspace.

Admins can govern AI usage centrally in Unity Gateway, including MCP policies, cost controls, and monitoring. MCP policies decide which MCP servers and tools Genie can use, which users can access them and which actions require human approval, and traces record what Genie did and who asked for it. Workspaces that use Private Link are supported through context-based ingress policies.

Questions about your company's data and knowledge don't come up in isolation, so they shouldn't be answered in isolation either. The Databricks Genie app for Slack brings governed answers and actions into the conversation where your team already makes decisions, with no new tool to learn and no new permissions model to manage.

Get started today:

To learn more about the Databricks Genie App for Slack, join our webinar: "Just Ask Your Data: Bring Databricks Genie into Slack"

Yes. Genie One is the default. It looks across your entire data estate to answer each question. You can also pin a conversation or channel to a specific Genie Agent. Questions are then sent to that agent through the Genie Agentic API, so answers stay scoped to that agent's curated data and instructions.

Genie in Slack does much more than answer questions. You can:

You have two options:

Either way, results arrive in your Genie DM. You need to sign in to the Genie app from a DM once before results can be delivered.

Yes. A Databricks account admin needs to create a context-based ingress policy to allow secure connectivity between Slack and your Databricks workspace. See the documentation for setup steps.

Yes. 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. The service that routes messages between Slack and Databricks doesn't store questions or answers.

Subscribe to our blog and get the latest posts delivered to your inbox.

── more in #ai-agents 4 stories · sorted by recency
── more on @databricks 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/genie-in-slack-is-no…] indexed:0 read:5min 2026-11-12 · —