Genie One expanded with a desktop app, Private Link support, and hybrid data grounding in Unity Catalog.

Genie received major enterprise workflow and infrastructure updates in August. Databricks introduced a dedicated Genie One desktop application featuring a global launcher, document collaboration tools, and expanded context powered by Genie Ontology snippets and file uploads [6]. For enterprise network security, inbound Private Link expanded to support account-level Genie One instances alongside the Databricks account console [18], [22]. Grounding capabilities also broadened as Databricks detailed architectures for running Genie Agents across both structured tables and unstructured Unity Catalog Volumes under user identity controls [26], [37], reinforced by curated Pages in Unity Catalog Semantics to supply verified citations [15].

Tooling around Genie Code gained traction for migration and automation tasks. Teams used Genie Code agent skills to modernize legacy data pipelines [23], [36] and automate the conversion of Power BI reports into Databricks AI BI dashboards with integrated Genie spaces [33], [39]. In addition, tutorials highlighted the Genie App Builder for deploying natural-language micro-applications tied directly to Unity Catalog data [40].

Practitioners tackled operational boundaries and external integrations. Developers ran into hurdles when configuring Genie Model Context Protocol (MCP) connectors within Claude workspaces [10] and Azure Copilot Studio [17]. Teams also debated data modeling strategies for text-to-SQL accuracy [16], [31], noted that image formats remained unsupported in beta volume attachments [27], and shared scripts to enforce hard cost controls and block Genie compute once budget caps were met [29], [30].