Alan Blythe

tech

Google Announces the Gemini Agent

At Gemini at Work 2026, Google Cloud announced the Gemini agent, a single agent for work that takes objectives, delegates to sub-agents, and runs on Gemini or Claude models.

2 min read

On October 8, 2026, at Gemini at Work 2026, Google Cloud CEO Thomas Kurian announced the Gemini agent, which Google describes as “a single, universal agent for work”.

What was announced

One agent, one API. In Google’s words, Gemini “answers your questions, handles your knowledge work, creates your images and media, and writes and runs code — all in a single agent and a single API.”

Objectives, not instructions. “You give it objectives, not instructions. You delegate an outcome and come back to finished work.” The agent plans the work, uses skills and tools, connects to company systems, and brings back the result inside the documents, inboxes and developer environments people already use.

Sub-agents and coworker agents. The agent can create temporary, job-specific sub-agents, each with its own identity, to handle multi-step work that can run for hours or days. It can also act as a coworker agent with a persistent role and its own Workspace account, including an email address.

Runs in the cloud. It keeps one set of memory and context whether you reach it from the web, mobile, desktop, the command line, Google Workspace, Microsoft 365 or Slack. It can also run headless inside third-party applications.

Connects to business systems. Google listed integrations including Gmail, Drive, Docs, Sheets, Slides, Chat and Calendar, plus tools like Slack, Jira, Confluence, BigQuery and Snowflake, and any Model Context Protocol (MCP) server.

Model choice. “Gemini is the agent, and the model underneath it is a separate choice.” It runs each job on the model that fits best, across Google’s Gemini models and Anthropic’s Claude models today, with other private and open models planned.

Cost and governance controls. Google announced routing to keep costs down, real-time spend caps, and enterprise controls: each agent gets its own identity and permissions, and its actions are recorded in an audit trail attributed to the agent.

Availability

The Gemini agent is in private preview for businesses, with availability coming to Google Workspace customers on select Business and Enterprise plans. Google says it will come to consumers later. Pricing wasn’t announced.

How it fits Google’s earlier approach

Analysts noted this is Google’s third enterprise AI positioning in under two years, after Agentspace and Gemini Enterprise. Futurum’s Nick Patience described the move as going “from somewhere to ask questions to somewhere to hand off work”, and pointed to two shifts:

  • From many agents to one. Rather than separate agents for separate jobs, Google is offering one agent with one memory and one set of controls.
  • The agent ahead of the model. Treating the underlying model as a swappable choice, including a competitor’s, is new for Google.

Coverage