Google launches Gemini Enterprise Agent Platform to build autonomous AI agents

Google Cloud has launched what it's calling an end-to-end solution for building and deploying autonomous AI agents at enterprise scale, and it open-sourced every line of demo code shown at the event.

The Gemini Enterprise Agent Platform, introduced at Google Cloud Next 2026, is the company's most structured attempt yet to move AI agents out of experimental prototypes and into production environments.

The platform spans the entire lifecycle of agent development, from an Agent Development Kit (ADK) for building agents to a serverless agent runtime for deploying and scaling them without infrastructure management.

Google Cloud President of Site Reliability Engineering Brad Calder described it as designed to help developers build agents that "proactively help users and complete tasks independently".

Beyond development and runtime, the platform introduces governance infrastructure: a unique agent identity per agent instance, an agent gateway for enforcing IAM policies, an agent registry for discovery, and an A2A protocol for agent-to-agent collaboration. Importantly, the platform supports models beyond Gemini Pro and Flash, including Anthropic's Claude, through Google Cloud's Model Garden.

To demonstrate the platform's capabilities, Google's developer keynote used a multi-agent simulation for planning a marathon in Las Vegas. The demo featured a planner agent, an evaluator, and a simulator working in coordination, drawing on mapping and GIS tools, stateful memory, RAG integration, and dynamic interfaces built at runtime.

Security was woven directly into the architecture rather than bolted on after. Agent Gateway acted as a proxy enforcing read-only policies on the finance MCP server, blocking the planner agent from executing write operations or accessing the open internet. The demo illustrated that governance isn't optional in this stack; it's structural.

Wiz Co-Founder Yinon Costica demonstrated how the platform integrates with cloud security tooling through a pair of specialised agents. A "Wiz Red Agent" identified an authentication bypass vulnerability running from an internet-exposed entry point to sensitive data.

A "Wiz Green Agent" then recommended prioritised fixes, downgrading IAM privileges, patching the bypass, and enforcing AI guardrails. Claude Code applied the changes, and Wiz rescanned to confirm resolution.

The sequence showed something important: AI agents handling not just task execution but also active security remediation within the same platform.

Google made the full demo codebase publicly available, including architectural guidance, labs, and developer credits. That move lowers the barrier for engineering teams wanting to test multi-agent workflows without rebuilding from scratch.