Google introduced a "universal" Gemini AI agent on Thursday that executes work tasks across applications and devices in the background, accessible through the Gemini Enterprise app. Google announced the tool at its Gemini at Work event and positioned it as a single chat interface where users assign tasks to the agent.

What surfaces does the agent reach?

The system operates inside Google Workspace applications including Gmail, Drive, Docs, Sheets, and Calendar. Beyond Google's own productivity suite, the agent connects to third-party platforms Slack and Microsoft 365. Interaction surfaces extend to mobile devices, desktop computers, the web, and the third-party apps themselves.

How does the cloud architecture change the use model?

Execution happens in the cloud rather than on the endpoint, which decouples the agent from any single session. The agent retains context — task state, prior instructions, intermediate outputs — as users move between devices or applications. That design enables asynchronous work: a user can assign a task from a phone, close the app, and retrieve results later from a desktop or web session.

Why does Google support competitors' stacks?

The cross-application scope is the central commercial bet. By including direct support for Microsoft 365 and Slack, Google positions the Gemini agent as workflow infrastructure rather than a Workspace-exclusive assistant. Enterprise IT shops rarely standardize on a single-vendor productivity stack; an agent that bridges Gmail, Outlook, Teams, and Slack addresses the integration gap that single-suite assistants leave open.

How is the product being sold?

Google distributes the product through the Gemini Enterprise app, the company's paid business tier. The announcement did not detail pricing, regional availability, or a specific general-availability date. The Gemini at Work event, hosted on Thursday, marks the formal product reveal but leaves the rollout schedule and license structure to subsequent communications.

Where does this fit in the enterprise AI race?

The launch adds Google to a competitive set that already includes Microsoft's Copilot inside Microsoft 365, Salesforce's Agentforce, and OpenAI's ChatGPT with its connector ecosystem. Differentiation, on Google's framing, rests on continuity of context across applications and devices rather than on depth of features inside any single product.

What does "universal" actually signal?

The label points to an architectural shift away from per-app assistants. Earlier generations of AI helpers in productivity software live inside a single application: a writing assistant in Docs, a mail helper in Gmail. Google's pitch collapses those into a single agent that crosses application boundaries, holding a task rather than a chat thread. Enterprise buyers evaluating Gemini against Copilot or ChatGPT will now compare agent ecosystems rather than individual features.

How does the Gemini Enterprise context shape expectations?

Google has built out Gemini Enterprise as a paid, business-focused tier distinct from the consumer Gemini product. The agent announcement extends that commercial lane: customers who already license Gemini Enterprise receive a workflow automation surface layered on top of the underlying model. The pricing structure, integration roadmap, and data-handling guarantees that typically accompany an enterprise AI product were not detailed at the event, leaving the procurement story for later disclosure.

The next signal to watch is whether Google publishes enterprise pricing and a general-availability date, and how Microsoft and Salesforce respond with their own cross-suite agent claims.