Google Cloud announced the Gemini agent, a universal enterprise agent that can perform knowledge work, content creation, coding and more from a single prompt box.
The Gemini agent connects to a company’s business systems, selects the best model for each task, and includes built‑in cost controls and governance features.
Available soon to Workspace customers on select plans, the Gemini agent can operate across Google apps, third‑party services and be invoked via familiar prompts in Gmail, Docs, Sheets, Slides and Chat.
body_html
Quick read · 1 min
Google Cloud unveiled the Gemini agent, a universal enterprise assistant that starts work from a prompt window and can handle questions, knowledge work, content creation, and coding across Workspace and third‑party services. It uses the best model for each task and includes governance and cost controls.
In practice, you could ask Gemini to draft a slide deck from a project update and automatically assemble the right participants’ calendars and emails. It acts as a team member or project manager, depending on how you configure it. Availability is coming soon to Workspace customers on select plans, with a private preview underway.
Google Cloud introduced the Gemini agent today at its Gemini at Work event, presenting a new universal assistant designed for enterprise users. The Gemini agent exists to start work directly from a prompt window and can handle tasks that range from answering questions to content creation to coding across a company’s tools and data.
In short, you tell Gemini what you want to accomplish, and it figures out the best path, the right models to use, and how to connect to your organization’s systems. It can plan work, use embedded skills and tools, and deliver finished results inside documents, inboxes, and developers’ environments you already rely on.
Google emphasizes security and governance as core: the agent is designed with admin controls and cost management so organizations can keep budgets in check and maintain oversight over AI actions.
01
What Gemini agent is and how it works
The Gemini agent is described as a single, universal agent for work that can be deployed within the Gemini Enterprise app, Workspace, and through third‑party services. Google says you can give it objectives rather than explicit instructions. It can do a broad set of tasks, including answering questions, doing knowledge work, creating media, and writing or running code.
Behind the scenes, the agent runs tasks on the model that fits best. Today, that can include Google’s Gemini models and Claude family models, with future plans to support other private and open models. It uses sub‑agents to handle multi‑step tasks and can coordinate workflows across tools and services.
02
Memory and how it keeps context
Gemini agent maintains a memory graph to stay productive over long tasks. There are four types of memory: session memory for the task at hand, semantic memory that builds a knowledge base from documents and conversations, procedural memory for how jobs get done, and episodic memory for everything it has done before. This setup lets the agent keep context across hours or even days of work.
03
Where you’ll see it first
The Gemini agent will be available soon for Workspace customers on select Business and Enterprise plans. It’s currently in private preview. The tool can be invoked from Gmail, Google Docs, Sheets, Slides, and Chat, and Google points to an API that will let developers surface the agent inside third‑party apps. The Gemini Enterprise app and Workspace Intelligence will be the main entry points for teams to start using it.
04
What this means for everyday users
For non‑technical workers, Gemini is pitched as a way to automate routine tasks and coordinate multiple people and tools without manual handoffs. Think of asking it to draft a project update in a slide deck, then automatically coordinating calendars and email threads to assemble the right participants. It can also act as a team member, working on behalf of a group or a specific role within an organization.
Google hasn’t shared a full rollout date yet beyond “soon” for Workspace customers on select plans. If you’re in a company that’s signed up for Gemini, you’ll likely see prompts to start enabling the agent in your workspace and an admin console to configure governance, costs, and access controls.
06
Quick answers
What is the Gemini agent?
A universal enterprise assistant that can handle knowledge work, content creation, coding and more across Google Workspace and connected apps, with an API for third‑party integrations.
When will it be available?
Google says it will be available soon for Workspace customers on select Business and Enterprise plans, currently in private preview.
How does it handle security and costs?
Google highlights built‑in governance and cost controls as core features, designed to keep enterprise usage safe and within budget.
