aiqbench takeAI Products

Google Cloud's new Gemini agent marks a significant shift toward autonomous execution within enterprise ecosystems, allowing models to dynamically invoke internal tools across Google Workspace applications without manual intervention. This development highlights the maturation of agentic workflows where AI can independently manage complex, multi-step tasks like data engineering and legal analysis while maintaining strict governance controls. The ability to switch between Gemini and Claude models further underscores a trend toward flexible, model-agnostic infrastructure for large-scale business operations.

Source covering36Kr

What this isA model's reading of the article and related coverage, written for aiqbench. Not the publisher's words.

Google Cloud Launches Gemini AI Agent Capable of Autonomously Executing Multi-Step Enterprise Tasks

36KrOctober 8, 20261 min read

Google Cloud announced the launch of Gemini, a general-purpose AI agent designed for enterprise workflows. Through a single interface, it can handle tasks such as Q&A, knowledge work, content creation, and coding execution while accessing internal tools and data to autonomously achieve user-defined goals. The agent integrates directly with Gmail, Drive, Docs, Sheets, Slides, Chat, and Calendar, and supports the creation of 'colleague agents' with independent identities and permissions. Additionally, Google introduced new skills for data analysis, data engineering, and machine learning, offering specialized versions for financial services and legal industries for the first time. Gemini enables dynamic switching between its own models and Anthropic's Claude, while providing enterprise-grade controls including permission management, auditing, sandboxing, and real-time spending limits.

Read the full article at 36Kr

Read at 36Kr ↗Ask about this article