On October 8, Google Cloud walked into the busiest room in tech and put down a single product: the Gemini agent, billed as one AI agent for work that answers questions, handles tasks, creates content, and writes code. The timing isn’t subtle. Every major tech company is racing to sell AI that works on its own, and Google wants the version that lives inside your workday to be theirs.
The agent plans the work, uses tools, connects to a company’s business systems, and returns finished work in documents, email, and developer environments. That’s the pitch. What makes it different is where it shows up: directly inside Google Workspace apps like Gmail, Docs, Sheets, and Calendar – and also through Microsoft 365 and Slack. Not a new tab. Not a new app. The place your day already happens.
What it actually does
The core idea is straightforward. Instead of a chatbot that answers one question at a time, this is an agent that takes a job and runs with it. Ask it to prepare a client brief and it can pull context from your inbox and calendar, draft the document in Docs, and hand back something finished rather than a starting point.
Two details stand out. First, it picks the best model for each task. It runs on Google’s own Gemini models and Anthropic’s Claude models today, with more to come. That’s an unusually open stance for Google: the agent isn’t locked to Gemini, it’s model-agnostic by design. Second, users can create “coworker agents” that act as team members – with their own email address and access only to the information they’re given. An agent that can be emailed, cc’d, and looped into threads like a colleague is an agent that fits into how offices actually run.
Google is also shipping specialized editions: financial services and legal are in preview now, with government, healthcare, and retail versions on the way. That’s the enterprise playbook in one sentence – sell the general agent, then tailor it for the industries with the biggest budgets and the strictest rules.
The bet everyone’s making
Here’s the thing about workplace AI. The last two years produced a pile of impressive tools that employees opened once and forgot. The pattern that keeps winning is different: AI shows up where people already work, quietly, as a feature rather than a destination.
Google’s move is the clearest expression of that thesis yet, and it’s not alone. This same week saw agents embedding inside Webex meetings on the same theory – the agent should arrive where the meeting already happens. The industry has converged on a conclusion: nobody wants another app to learn. Everybody wants the apps they have to get smarter.
There’s quiet logic in the model-agnostic choice, too. If your agent’s job is to pick the best model per task, Google is admitting something honest: no single model wins everything. For an enterprise buyer worried about lock-in, that’s a better pitch than “ours is best.” For Google, it keeps the agent competitive even in weeks when someone else’s model scores higher. Pragmatism as strategy.
What it means for your workweek
The inbox becomes a command line
The “coworker agent” concept deserves a second look. Giving an agent its own email address means anyone in the company can assign it work with a cc – no training required. That’s a lower bar than any AI rollout so far. The risk is obvious: agents with email addresses can also be phished, spammed, and confused by thread context. Google’s answer is scoped access – only the information it’s given. Whether that’s enough will be decided by the first embarrassing incident, and there will be one.
Governance, not procurement
The question for managers is shifting. It used to be “which AI tool should we buy.” It’s becoming “which agents do we let loose in our systems.” That’s a governance conversation, and it’s the one every CIO is already having. Google’s pitch is that an agent with clear access boundaries and an audit trail is easier to govern than a dozen shadow tools employees pasted API keys into.
Watch the industry editions
Financial services and legal first, then government, healthcare, retail. That order tells you where Google thinks the money and the tolerance for AI assistance overlap. If the legal edition survives contact with actual law firms – with their very particular relationship to accuracy – the enterprise agent race gets a lot more interesting.
Strip away the launch theater and there’s a genuinely encouraging trend. The AI industry spent years building impressive things that demanded people change their habits. Now it’s building things that slot into existing habits. The chatbot era was the awkward phase: clever answers, no follow-through. The agent era is the part where the follow-through arrives. The best AI assistant for work won’t be the one you visit. It’ll be the one that’s already there.
Be First to Comment