Databricks expands Genie One into an enterprise AI cowork home

Search, directories and finished work in one place

Genie One brings together tools for finding information, connecting with colleagues, creating work and taking action. Its new Genie Search combines context-aware chat, live information retrieval and enterprise search.

According to Databricks, Genie Search indexes organizational knowledge held outside Databricks, including documents, tickets and wikis in Google Drive, Jira, Confluence, Workday and SharePoint. It also supports over 100 MCP connectors, allowing Genie to retrieve live information and take actions such as checking or creating a Jira ticket. Genie chooses between search, MCP and SQL based on the task.

The same index supports asset search, allowing users to find external documents alongside Databricks dashboards, tables and Genie Agents. Genie One also includes organizational directories, where users can browse team structures, find colleagues and understand reporting relationships. Admins and users can pin dashboards, apps and Genie Agents to the homepage.

Genie One can create artifacts directly in a conversation, including slide decks, dashboards and interactive pages. Users can refine these outputs with Genie, inspect the sources behind numbers or visualizations, and make changes before sharing. Scheduled tasks can keep artifact data current, while skills can apply company brand standards.

From answers to actions across workplace tools

The release extends Genie One beyond generating information. Through MCP connections, Genie can read and update systems such as Jira and Asana, allowing the results of an analysis to reach a system of record without leaving the conversation. Databricks gives the example of asking Genie to summarize open escalations in Jira and post themes to a Slack channel.

Users can choose which actions require approval, while administrators can decide which tools their teams may connect to Genie. Skills allow employees to capture a workflow once and share it with colleagues. Published to Unity Catalog, each skill runs with the permissions of the person using it, so access remains limited to that user's authorized data and tools.

Databricks also says teams can bring skills created in other AI tools into Genie One. The company describes the product as an always-on coworker for workflows across teams and tools, with conversations, artifacts and workflows drawing on enterprise context from Genie Ontology.

Access across devices and collaboration platforms

Genie One is available through a browser extension for Chrome, native desktop apps for macOS and Windows, and mobile apps for iOS and Android. The desktop apps support questions about data, threads and documents, Genie Agents, Unity Catalog tables and Databricks Apps. Mobile users can ask questions, explore and chat with dashboards, and run Databricks Apps.

Databricks is also bringing Genie One into Slack and Microsoft Teams. Users can message Genie directly or mention it in a channel or group chat, while shared answers can give teammates context to build on. Channel owners can pin a Genie Agent for a specific use case. Genie preserves each user's identity and permissions across these experiences, according to Databricks.

For organizations, the expansion means Genie One is intended to serve as a common layer across search, analysis and action rather than another assistant limited to a single application. Unity Gateway provides administrators with traces, MCP policies and budget controls for managing Genie as adoption grows. The company says Genie plugins also bring Genie into ChatGPT, Claude, Microsoft Copilot Cowork and Grok Bot.