Google Enterprise Agents

Google used Cloud Next on April 22 to turn Vertex AI into the Gemini Enterprise Agent Platform, a new system for building and managing fleets of artificial intelligence agents. (cloud.google.com) An AI agent is software that can plan steps, call tools, and complete tasks with limited human input; Google said the new platform adds orchestration, security, DevOps, and optimization on top of Vertex AI’s model-building tools. (cloud.google.com) Google also added Agent Studio, a low-code workspace for building and testing agents, alongside documentation for connecting outside agents into Gemini Enterprise through the Agent2Agent, or A2A, protocol. (cloud.google.com) (docs.cloud.google.com) A2A is Google’s open protocol for agent-to-agent communication, which lets one agent discover another, hand off work, and exchange updates across different platforms. Google said A2A complements Anthropic’s Model Context Protocol, which focuses on giving agents access to tools and context. (developers.googleblog.com) Google has been laying the groundwork for months. In 2025, it introduced A2A and later said it had donated the project to the Linux Foundation with backing from Amazon Web Services, Cisco, Microsoft, Salesforce, SAP, and ServiceNow. (developers.googleblog.com 1) (developers.googleblog.com 2) The company tied the launch to a broader push for what it calls the “agentic enterprise,” where businesses run multiple specialized bots instead of a single chatbot. Sundar Pichai wrote in Google’s Cloud Next roundup that the event’s roadmap stretched from the new agent platform to custom chips and other cloud infrastructure. (blog.google) Google’s pitch is that companies want one governed stack as they move from experiments to production deployments. Thomas Kurian said the platform is designed to “build, scale, govern, and optimize” agents, while Google’s documentation says it is meant for enterprise data, applications, and workflows. (cloud.google.com) (docs.cloud.google.com) The competitive backdrop is crowded. Microsoft, Amazon, Anthropic, Salesforce, and ServiceNow are all pushing agent frameworks and standards, and Google’s answer is to bundle models, tooling, interoperability, and cloud operations into one product line. (developers.googleblog.com) (zdnet.com) The change also clears up Google’s branding. ZDNET reported that Google had “reworked Gemini Enterprise” for this launch, but Google’s own announcement describes the move more directly as an evolution of Vertex AI into a platform for enterprise agents. (zdnet.com) (cloud.google.com) Google is no longer selling enterprises on a single model or chatbot. It is selling the control tower for a growing number of bots that need to work together without breaking company rules. (cloud.google.com)