Google's New Gemini Agent Gets Its Own Email Address

Google has launched what it calls a universal agent for work. According to Google's announcement and TechCrunch's report, the Gemini agent doesn't just answer questions. You give it a goal, and it plans the work, uses tools, and hands back finished output in the documents and inbox you already use.

What actually happened

Google introduced the Gemini agent on 8 October at its Gemini at Work event. A few details from the Google Cloud write-up and TechCrunch:

• It works inside Gmail, Drive, Docs, Slides, Sheets, Chat and Calendar, and is also reachable from Microsoft 365 and Slack.
• A "coworker" agent can be given its own Workspace account, with an email address, calendar and Drive. It acts under its own identity rather than yours and, Google says, sees only what is shared with it.
• It can set up temporary sub-agents for multi-step jobs, and Google says some jobs can run for hours or days.
• Google says every action goes into an audit trail attributed to the agent, and that administrators approve agent permissions.
• It chooses a model for each task. Gemini and Anthropic's Claude models are available at launch.
• Businesses come first. TechCrunch reports consumers will get it later, with no date given.

Google has not published agent pricing. The announcement also doesn't say which Workspace plans include it or whether the core agent is generally available or still in preview.

What it means for a small business or recruiter

Until now most people used AI by typing into a chat box and copying the answer out. This is a step toward delegating: "pull together last month's candidate replies and draft follow-ups" rather than "write me an email."

The detail worth noticing is the separate identity. An agent with its own email address and Drive is closer to a new team member than a tool. That is useful, because you can see what it did and limit what it can reach. It also means you need to treat it like a new hire when it comes to access and sign-off.

What to do

1. Don't switch anything on yet. Details on plans and pricing aren't published. Wait until you know what it costs and what it can touch.
2. Tidy your sharing first. Google says the agent only sees what is shared with it, so messy folder permissions become your problem. Fix who can see client and candidate files now.
3. Start with low-risk jobs. Summaries, first drafts and research are fine. Anything that sends, deletes or pays should need a person's approval.
4. Ask who reads the log. An audit trail only helps if someone looks at it. Pick a person.
5. Keep candidate and client data in mind. If you hold CVs or client details, check your data-protection obligations before an agent gets near them.

The caveat

This is a vendor announcement. The productivity claims are Google's and customer results are vendor-reported. Several things that matter to a small business, including price, plan availability, rollout timing and how the admin controls actually work, aren't stated yet. TechCrunch lists early testers and large enterprise customers, not small firms, so how it performs for a ten-person agency is unknown. Features also tend to arrive in stages, so what you can use this month may be less than the keynote suggests.

If you're weighing up where AI agents fit in your recruitment or sales process, talk to us.