Home / Technology / Google Launches Agent Platform for AI Workflows
Google Launches Agent Platform for AI Workflows
22 Apr
Summary
- Gemini Enterprise Agent Platform integrates AI model selection and building.
- The platform offers over 200 models, including Gemini, Gemma, and Anthropic.
- Enhanced security features like Agent Identity and Agent Simulation are included.
Google has launched the Gemini Enterprise Agent Platform, an advanced solution built upon Vertex AI to manage automated agents in business workflows. This platform aims to address the increasing challenge of securely and efficiently managing numerous agents.
It consolidates model selection, building, and tuning services from Vertex AI, while introducing new capabilities for agent integration, security, DevOps, and orchestration. The platform provides access to over 200 models, including Google's own Gemini and Gemma, alongside competitive offerings from Anthropic.
Developers can manage the agent lifecycle from design to deployment and governance. Features such as MCP support, an upgraded Agent Development Kit, faster runtime, and Memory Bank enhance agent reasoning and context handling for complex tasks.
Security is a core focus, with tools like Agent Identity providing cryptographic IDs for each agent and Agent Simulation allowing for pre-deployment stress testing against real-world scenarios.
Published agents are accessible via the Gemini Enterprise app, supporting no-code or lower-code development for employees. Google emphasizes enhanced governance and security, ensuring IT visibility and auditability comparable to essential business applications.
Additionally, Google introduced Agentic Data Cloud for scaling AI agents and Workspace Intelligence, which uses Gemini reasoning to understand complex relationships within Workspace apps for enhanced automation.
