Google unveils the Gemini agent for enterprise in private preview

Google unveils a Gemini agent designed to automate professional tasks across its tools and third-party services. The launch begins in private preview for enterprise customers via the Gemini Enterprise application. The tool ensures continuity of context across devices and offers a dedicated agent identity to collaborate like a colleague.

Limited Access for Enterprises and Private Preview Testing

The Gemini agent is currently reserved for enterprise customers in private preview. Google already offers a public version with Gemini Spark, launched in May, which allows subscribers to perform tasks in Workspace and on the web. This public version is accessible through a separate interface in the Gemini application. The enterprise-focused agent could signal Google’s intention to provide a centralized AI assistant capable of answering questions and executing tasks from a single window.

A Unified Interface and Multichannel Integrations

The Gemini agent will be available in the Gemini Enterprise application, where users can interact with it and assign tasks from a single interface. It will be accessible from mobile devices, desktops, the web, as well as third-party applications like Slack or Microsoft 365. For internal tools, it will be able to operate directly within Gmail, Drive, Docs, Sheets, and Calendar. Operating in the cloud, the agent will maintain the same context across all connected devices.

"Coworker" Agent Functions and Specialized Sub-Agents

Presented as a universal agent operating in the background across various applications and devices, Gemini can collaborate with specialized sub-agents depending on the task at hand. It can also function as a coworker agent, equipped with a dedicated identity and an email address in the format “@agents.company.com.” Google specifies that the agent will automatically execute tasks using the most suitable AI model. The announcement took place during the Gemini at Work event.