AI · Oct 9, 2026
The Gemini agent in Google Workspace: what Google Workspace businesses should know
At its Gemini at Work event on October 8, Google Cloud introduced the Gemini agent. Google Cloud CEO Thomas Kurian described it as a single, universal agent for work that answers questions, handles knowledge work, creates images and media, and writes and runs code, all from one prompt box.
If your business runs on Gmail, Drive, and Google Calendar, this is the AI that is heading for your team's inbox. Here is what stood out.
It works inside the apps you already use
Google says the Gemini agent works inline in Gmail, Drive, Docs, Slides, Sheets, Chat, and Calendar, carrying the same memory, skills, and controls everywhere. It also runs on web, mobile, and desktop, and can be reached from Slack and Microsoft 365.
Google gave a simple example. You ask Gemini to set up a meeting with "the usual team of regional event leads" next week, without naming anyone. It works out who those people are from your chat space and past email threads, checks calendars, and starts an email to find a time.
That kind of example matters because it reflects how small businesses actually work. Nobody has time to write detailed instructions. The useful AI is the one that already knows the context.
Coworker agents with their own accounts
The most interesting piece for growing teams is what Google calls a coworker agent. You describe a role, and Gemini creates an agent with its own Workspace account, including an email address, calendar, Drive, and a spot in your company directory. Colleagues work with it by adding it to a Chat space or @mentioning it.
Google says a coworker agent acts under its own identity rather than yours, sees only what you share with it, and appears under its own name in document version history. That is a sensible design. When something goes wrong, you can see exactly which agent did what.
Cost controls are built in
Google spent a good part of the keynote on cost. It described three controls:
- Multi-model orchestration, so each job runs on a model suited to it. Google says it orchestrates across its own Gemini models and Anthropic's Claude models today.
- Smart Routing, which automatically sends each workload to the model that gives the best performance at the lowest cost.
- Real-time spend caps, where you set a hard limit on a project's AI spend and the agent pauses when it hits that limit.
Spend caps are the feature I would turn on first. Background agents can use a lot of compute, and a hard limit is the simplest protection against a surprise bill.
A South Carolina example
Among the customers Google highlighted was Santee Cooper, South Carolina's state-owned electric and water utility. Google says it is using a specialized model built on Gemini Enterprise for financial scenario modeling. It is a reminder that this is not only for Silicon Valley companies. Organizations in our region are already putting it to work.
What to do before it reaches your team
- Decide who gets access first. Start with a few people who will use it daily and report back.
- Clean up sharing. Gemini works with what your people can access. If old Drive folders are shared too broadly, the agent can see them too.
- Set spending limits. Turn on spend caps before anyone sets up a long-running task.
- Name an owner for each coworker agent. Someone should be responsible for what it does, just like a new hire.
What this means for your business
- If you use Google Workspace, an agent that works inside your email, documents, and calendar is coming. Plan for it now.
- Coworker agents with their own identities make it easier to track and limit what AI does.
- Use spend caps from day one.
- Tidy up file sharing permissions. They decide what the agent can see.
Whether your business runs on Google Workspace or Microsoft 365, we help teams roll out AI with the right permissions and spending limits in place.
