AIToday
Large Language ModelsAI Business & IndustryGIGAZINE AIPublished: Oct 9, 2026, 16:00 JST

Google unveils Gemini agent for cross-app work tasks

Google unveils Gemini agent for cross-app work tasks

3 Key Points

  1. What happened

    At Gemini at Work, Google introduced the Gemini agent, built into Gemini Enterprise, which gathers information and plans its own steps across Google Workspace, Microsoft 365, Slack, and Salesforce.

  2. Why it matters

    Employees who today hand information between email, documents, spreadsheets, and internal systems could hand a whole goal to one agent instead, though it is a limited preview for some companies.

WHO IT HITSEnterprise IT and operations teams that manage workplace software and access permissions will need to get comfortable with agents that log their own activity and carry separate identities. Knowledge workers in finance and legal are the first targeted users, with government, healthcare, and retail named as next.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Context & Analysis

Google already sells Gemini Enterprise, where employees search internal information or use task-specific AI agents. The problem it left unsolved was that company work runs across mail, documents, spreadsheets, and internal systems, so delegating a job often meant moving information and spelling out each step by hand. The new Gemini agent is meant to close that gap by planning and executing the steps itself, and it keeps working in the cloud even after a user closes their laptop.

The design goes beyond a single assistant. Users can create specialist agents, run several in parallel with different tasks, and stand up a "Coworker agent" that gets its own email address, calendar, and Drive, so staff can summon it in Google Chat or from a document comment. Google is also pairing the launch with industry versions for finance and legal work.

Control features appear aimed at the concerns that come with that reach. Agents get their own identities with logged activity and adjustable permissions, and external communication passes through a mechanism called Agent Gateway so company security rules apply. Whether these safeguards satisfy IT teams may shape how quickly the agent moves past its limited preview.

FAQ
Can I use the Gemini agent today?
No. It is available as a limited preview for some companies only, and Google has not announced when general availability starts.
Which AI models does it use?
It supports Google's Gemini series and Anthropic's Claude series, so users can pick the model that fits each task to aim for better accuracy at lower cost.
How does Google keep spending under control?
Spending limits can be set per project, and when an agent reaches the cap its work pauses.

Also reported by Impress Watch, TechCrunch AI

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleOpenAI posts 719 math proofs from unreleased model