Google Cloud has introduced Gemini Agent, a single AI agent meant to handle a wide range of enterprise work rather than requiring separate tools for separate tasks. The company unveiled it on 8 October at its Gemini at Work event, where chief executive Thomas Kurian described the approach as working from objectives rather than step-by-step instructions.
Each Gemini Agent gets its own Google Workspace account, including a Gmail address, calendar and Drive storage, so it can send email, schedule meetings and store files like a human colleague would. It can run scheduled jobs, react to events and keep working in the cloud for hours or days, even while the user who assigned the task is offline. For complex jobs, it can spin up temporary specialised sub-agents and coordinate them in parallel or in sequence.
Gemini Agent connects to Workspace apps as well as third-party tools, including Slack and Microsoft 365, and can route individual tasks across Google's own Gemini models or Anthropic's Claude models, with support for further models planned. Governance features include dedicated agent identities, role-based permissions, audit trails, sandboxed execution and an Agent Gateway that applies network and policy controls to what agents can access.
The product has entered private preview with a small group of enterprise customers, reportedly including On, Shopify and PayPal, with no date yet set for broader availability.
Source: MarkTechPost
