Incident sheet · /enterprise-ai

Filed
Read
4 min · 730 words
Severity
Routine

Databricks Genie Goes GA in Teams and Plugs Into Microsoft Copilot

Databricks Genie is now GA in Microsoft Teams, with a new plugin connecting Copilot and Copilot Cowork to Genie One, plus Excel, Copilot Studio and Foundry integrations under Unity Catalog governance.

What's new for Genie in Microsoft: Copilot Cowork and Chat, Microsoft Teams, and Excel.
FeedWhat's new for Genie in Microsoft: Copilot Cowork and Chat, Microsoft Teams, and Excel. — AI-generated

Event log

  1. The Databricks Genie app for Microsoft Teams is now generally available
  2. Genie Ontology, the enterprise context layer, is now generally available
  3. The Azure Databricks Genie plugin connects Copilot Chat and Copilot Cowork to Genie One with no OAuth registration required
  4. Unity Gateway gives admins MCP policies, traces, and budget controls across Microsoft tools
  5. Genie will soon connect to Work IQ over MCP under Microsoft Entra ID identity

The Databricks Genie app for Microsoft Teams is now generally available, and Databricks has shipped an Azure Databricks Genie plugin that connects Microsoft Copilot and the new Copilot Cowork to Genie One — putting governed lakehouse data inside the tools employees already open every day.

Most AI assistants can read your email, files and calendar. What they can't do is tell which revenue table is certified, or how finance defines net retention. That gap produces inconsistent answers, low adoption and distrust among business users. Genie Ontology, also now generally available, attacks the problem directly: it combines business semantics teams explicitly define — metric views, certified definitions — with context learned from dashboards, queries and documents, so every answer lands on a shared understanding of the business.

What does that change for Copilot users?

Instead of Copilot hunting for answers on its own, it can delegate to Genie One. Genie answers using certified definitions and metric views, all governed by Unity Catalog, so the numbers match what finance already reports.

Databricks sketches the workflow: an account director asks Copilot Cowork for a briefing before a customer meeting. Cowork delegates to Genie, which pulls usage, spend and account trends from Azure Databricks using certified metric definitions, then reads recent Outlook email threads through the Microsoft 365 connector to flag open escalations. Cowork turns that into a Word brief, an Excel chart pack and a client-ready deck — every figure sourced from Genie. Setup takes a few clicks: turn on the plugin, click Connect, sign in with Azure Databricks credentials. No OAuth app registration, no manifest edits.

How does Genie work inside Teams?

Users can direct-message Genie or type @Databricks Genie in any channel or group chat, and it replies in the thread with answers grounded in governed business context. Databricks gives a concrete example: a sales manager posts "Which open leads over $100K haven't had a follow-up this week?" — work that used to mean pulling the pipeline report, cross-checking the CRM, searching meeting notes and messaging reps one by one. Genie returns the matching leads, their latest activity and links to the sources it used.

Beyond answers, Genie in Teams lets users:

  • Pin a channel or conversation to a specific Genie Agent, with questions routed through the Genie Agentic API
  • Upload files, render charts and trigger MCP actions
  • Turn insights into work without leaving the Teams workspace

Genie One is the default — it searches across the whole data estate. Admins add the app, set default response visibility, and users sign in to Databricks and pick a workspace on first use. Workspaces using Private Link need a context-based ingress policy created by an account admin. Security holds per user: every question runs as the person who asked it, processing happens inside the customer's Databricks workspace, and the routing service between Teams and Databricks stores no questions or answers.

Databricks recommends the native Teams app over publishing a Copilot Studio agent to Teams — it chats directly with Genie One, works in Microsoft 365 Copilot, and needs no Copilot Studio license.

What about Excel and builders?

The Azure Databricks Excel Add-in now carries Genie One. Analysts ask questions in plain English and get answers from governed data without writing SQL. Every answer includes citations, and analysts can refine with follow-ups, resume past conversations and load table results directly into the workbook — no exports, no copy-paste reconciliation.

Builders can wire Genie Agents into Microsoft Copilot Studio through the Azure Databricks connector for Power Platform, with Unity Catalog governance on every answer. Genie also reaches Microsoft Agent 365 and Microsoft Foundry, where Foundry-built agents can act as the signed-in user against governed Databricks data, and Foundry models flow back into Azure Databricks through Unity Gateway.

Who's watching the budget?

Admins govern the whole footprint through Unity Gateway: MCP policies decide which servers and tools Genie can use, who can access them and which actions need human approval; traces record what Genie did and who asked; cost controls keep AI spend predictable. Unity Catalog enforces each user's data permissions on every request.

Coming soon: Genie will connect to Work IQ over MCP, pulling email, meetings, documents, Teams messages and organizational context into Genie alongside governed Databricks data — each request running under the signed-in user's Microsoft Entra ID identity, with sensitivity labels and compliance policies intact.

via databricks.com (Source)

Filed under

  • databricks-genie
  • microsoft-copilot
  • microsoft-teams
  • azure-databricks
  • unity-catalog
Share this article:

More from Olivia Hart

Olivia Hart

Show full bio

News editor covering money, funding and market shifts around enterprise information technology and the leaders who run at CIO Monitor.

207 articles

The newsroom digest

One email a day with the stories that matter. No noise.

By subscribing you agree to the privacy policy.

Cross-references

‹ Previous articleNext article ›