Google Cloud has announced the Gemini agent, which it describes as a “single, universal agent for work”. 9to5Google reports the launch came at the company’s Gemini at Work 2026 event, with the agent aimed at enterprise customers and available through the Gemini Enterprise app, Workspace and third-party services.
The pitch is delegation. Google says you can “give it objectives, not instructions”, and the agent then answers questions, handles knowledge work, creates media and writes and runs code. Google Cloud CEO Thomas Kurian put it this way in a blog post, as Constellation Research quotes: “You delegate an outcome and come back to finished work.”
How the agent works
One example from Google’s materials: a manager emails asking for a project update as a slide deck. Workspace Intelligence flags it as a delegatable task and offers a single-click handoff to Gemini. In another, you ask Gemini to set up a meeting with your regional event leads without naming anyone. It identifies them from the chat space and your last event thread, checks calendars, and starts an email thread to find a time, including external participants.
Beyond personal assistance, Google says the agent can act as a team member, either for a group, like a project manager, or for a role, like an analyst in a finance department. It routes each job to the best-fit model. Today that means the Gemini and Claude families, with other private and open models promised later, and Google says this is meant to lower costs. Constellation Research adds that the agent includes cost controls and is governed by enterprise policies.
It uses sub-agents for multi-step work, including parallel and sequential steps that can run for hours or days. The agent runs in the cloud with one set of memories and one personalisation graph, split into four types:
- Session memory: the task in hand, even when it runs for days.
- Semantic memory: a structured knowledge base built from documents, conversations and other agents.
- Procedural memory: how a job gets done, including skills the agent writes for itself.
- Episodic memory: everything it has done before.
You can invoke it by mentioning @Gemini in Gmail, Docs, Sheets, Slides and Chat. You can also reach it from the command line, Slack, and Microsoft 365, and an API lets third-party apps use it as a “headless agent” with no dedicated interface.
Availability
The Gemini agent is in private preview. Google says it will be widely available soon for Workspace customers on select Business and Enterprise plans. The sources don’t give pricing, a date or a list of supported regions, so there is no word yet on when UAE businesses can use it.
The launch continues Google’s push to put Gemini inside everyday work tools, following the Gmail AI Inbox plans and the Gemini 4 Argon rollout. Using Claude alongside Gemini is also a notable choice, since Grok Bot has also moved to Claude.
What is the Gemini agent?
It is a Google Cloud agent for enterprise work that takes an objective, plans the steps, uses tools and sub-agents, and returns finished work across Workspace and connected business systems.
Which models does it use?
It picks a model per job. At launch, that covers Google’s Gemini and Anthropic’s Claude families, with other private and open models planned.
Can I use it now?
Not yet. It is in private preview, and Google says Workspace customers on select Business and Enterprise plans will get wider access soon.
