Databricks Upgrades Genie One with Desktop App and Agents

Databricks has introduced a suite of new features for its Genie One assistant designed to reduce context-switching and accelerate post-analysis workflows. Chief among these updates is the launch of a native Genie One desktop app, currently available in beta for macOS. The desktop application features a global launcher and maintains persistent conversations with the web version, allowing users to seamlessly transition between environments. Additionally, Databricks has enhanced document collaboration within the platform, enabling users to draft, edit, comment on, and share reports containing live data visualizations directly inside Genie One. Users can also download these documents as PDFs or export them to external tools.

To help teams replicate and scale their workflows, Genie One now allows users to convert active chat conversations into reusable Genie Agents, provided they have Workspace or Databricks SQL access. These agents retain domain-specific context and can be shared with colleagues to ensure consistent, governed data analysis. For one-off tasks, users can generate read-only links to share entire chat histories, including questions, answers, and visualizations, with other account users without requiring them to repeat the analytical work.

Databricks is also expanding the data context available to Genie One. The assistant now utilizes Genie Ontology snippets, which automatically extract business rules and metric definitions from dashboards, SQL queries, Metric Views, Domains, Pages, and existing agents. Users can also upload CSV, Excel, PDF, image, and Word files directly into private chat sessions for immediate analysis. Furthermore, Model Context Protocol write actions allow Genie One to perform tasks in external tools, such as drafting emails or updating tickets. Security and access permissions for these integrations are managed through the Unity AI Gateway, while Databricks data and Genie Agents remain governed by Unity Catalog. External integrations with platforms like Google Drive, SharePoint, GitHub, Glean, and Atlassian rely on their respective user authentication protocols.