Google has launched Gemini agent — a new enterprise AI assistant that works across Gmail, Docs, Sheets, Slides, and Chat as a persistent, memory-aware team member. It’s not a chatbot. It’s designed to accept high-level goals (“Prepare Q3 budget review for leadership”) and execute multi-step workflows without step-by-step prompting, coordinating sub-agents, running code, pulling data from connected tools, and even managing long-running tasks over hours or days. The feature is rolling out now in private preview for Google Workspace Business and Enterprise customers. Wider availability is expected “soon”, but no public date or pricing has been announced. It will be embedded in the Gemini Enterprise app, Workspace web and mobile interfaces, Slack, Microsoft 365, and via API for headless integration. Command-line access is also supported. The agent dynamically routes work across Gemini and Claude models (and, later, private or open models) to balance quality and cost. A shift from fixed-model deployments. Memory is unified and cloud-hosted: one context graph, one personalization layer, one set of memories across all channels. Mentioning @Gemini in supported apps triggers it instantly. This is not a consumer product. There’s no mention of Gemini agent appearing in the free or personal-tier Gemini apps, Pixel devices, or Android. It lives strictly in the enterprise stack, and it’s built to replace human-coordinated handoffs between tools, teams, and timelines. Next step: public rollout to eligible Workspace plans.