Google Cloud has unveiled Gemini agent, a new AI-powered tool designed for corporate users. The agent is integrated into Gemini Enterprise, the Workspace suite, and other services. It is capable of handling complex, multi-step tasks that can span several hours or even days, requiring only high-level goals rather than step-by-step instructions.
The agent functions by selecting appropriate models, including Google Gemini and Anthropic Claude, with plans to add more proprietary and open-source options. It utilizes sub-agents for complex workflows and interacts with various third-party tools. The system operates in the cloud, leveraging a unified memory, context, and personalization graph to maintain continuity.
Users can invoke the agent by mentioning @Gemini in Gmail, Google Docs, Sheets, Slides, or Chat. It is accessible across multiple platforms, including Slack, Microsoft 365, and via API for headless applications. Currently, the product is in closed beta testing and will eventually be rolled out to specific Google Workspace Business and Enterprise plan subscribers.
