How Google just revamped Gemini Enterprise for the agentic era - here's what's new

Google has fundamentally retooled its enterprise AI offering for what it calls the 'agentic era,' launching the Gemini Enterprise Agent Platform at its Cloud Next conference. Evolved from Vertex AI, the platform is designed as a comprehensive solution for companies deploying fleets of autonomous AI agents. It consolidates model selection, building, and tuning with new features for integration, security, DevOps, and orchestration. A key technical upgrade is support for over 200 models, including Google's own Gemini 3.1 Pro and Nano Banana 2, open Gemma models, and competitive offerings like Anthropic's newly released Opus 4.7. All Vertex AI services will now flow exclusively through this new Agent Platform.

The platform enables developers to manage an agent's entire lifecycle, from creation to scaling and governance. Enhanced reasoning is facilitated through an upgraded Agent Development Kit and MCP support, which structures agents into sub-networks for handling complex, multi-step workflows. Security is a core focus, addressed by features like Agent Identity, which assigns each agent a unique cryptographic ID, and an Agent Simulation tool for stress-testing before deployment. Once built, agents can be published to the Gemini Enterprise app, where business users can deploy teams of agents using no-code tools like Agent Studio to tackle projects—demonstrated in a use case where multiple agents collaborated on a furniture company's inventory and marketing strategy using Workspace data.