Google brings agentic AI to Gemini, starting with businesses

Your AI assistant can now take action on your behalf, and you can track its work in real-time.

Google has announced the integration of agentic capabilities into its Gemini AI, beginning with a unified agent designed for businesses. This new agent can perform tasks on behalf of users from a single interface, moving beyond conversational AI to execute actions like scheduling, booking, and code generation. The move aligns with the broader industry trend of AI tools becoming more task-oriented and capable of taking ownership. Google CEO Sundar Pichai highlighted Gemini's existing user base of over 1 billion monthly active users, with nearly 90% of Fortune 100 companies already using Gemini Enterprise. Initially targeting businesses, the agent will later be rolled out to consumers. Google Cloud CEO Thomas Kurian explained that the agent can be assigned "objectives, not just instructions," enabling it to plan, utilize custom tools, and connect to internal business systems. It will support various models, including third-party options like Anthropic's Claude, with plans to expand to open-source and private models. The agent can securely access company data and systems, functioning like a coworker with its own Workspace account, email, and awareness of internal organizational structures. Users can interact with it via tagging, email, or group chats, and track its actions through a dedicated "tasks inbox."

Google's strategic push to embed agentic AI within Gemini, prioritizing enterprise clients, signals a significant market shift towards AI systems that actively execute tasks rather than merely responding to queries. The company's vast existing user base and strong corporate adoption provide a robust foundation for scaling these agentic capabilities, potentially setting a new standard for workplace AI integration. This focus on "objectives, not just instructions" and the ability to connect with a wide array of business systems suggest a move towards more autonomous and integrated AI workflows within organizations, aiming to solve complex business problems through AI collaboration.

The introduction of an agent with its own identity and access to sensitive company data raises critical considerations around security, privacy, and governance, which Google aims to address through advanced security protocols and granular control features. By allowing users to specify models and providing transparent task tracking, Google is attempting to balance powerful automation with user trust and control, a crucial element for widespread adoption in the enterprise. The success of this initiative could compel competitors to accelerate their own agentic AI development and integration strategies, further intensifying the race to deliver AI solutions that augment human capabilities by performing discrete actions.