At a Google Cloud event held on Thursday, Google announced it is bringing Gemini AI into the agentic era, launching a unified agentic system that achieves the leap from "answering questions" to "executing tasks on users' behalf."
Currently, Gemini has more than 1 billion monthly active users, and nearly 90% of Fortune 100 companies have adopted Gemini Enterprise. Given the extremely high requirements of enterprise scenarios for security, scale, and performance, Google chose to deploy this agentic service on the enterprise side first, before expanding it to the consumer market.
The new system has goal-oriented planning capabilities, supports custom skills and tools, can deeply connect with internal enterprise data and systems such as Google Workspace, Microsoft 365, Slack, and Jira, and securely supports various Model Context Protocol (MCP) servers inside and outside the network. The system matches the best model by default, while also offering choices of third-party and open-source models, including Anthropic Claude.
Notably, the AI agent has its own Workspace account and email address, allowing it to act as a "digital colleague" integrated into team collaboration and process approvals. Companies such as On, Shopify, and PayPal are already participating in early testing.
Generative AI is currently accelerating its evolution from a single conversational interface toward automated task execution. By embedding agents throughout the enterprise ecosystem, Google not only strengthens its competitive moat against giants like Microsoft in the B2B market, but also marks the formal entry of enterprise AI applications into a new stage of advanced autonomy and human-machine collaboration.