Google launches Gemini Enterprise — no-code platform for long-running email-automation agents

Google Cloud used Next ’26 to launch Gemini Enterprise Agent Platform, a new system for building, deploying, governing and optimizing business AI agents. (cloud.google.com) (blog.google) The product combines Vertex AI’s model-building tools with new agent integration, security and DevOps features in what Google called a “single destination” for technical teams. (cloud.google.com) (blog.google) Google is also pushing the platform through the Gemini Enterprise app, where employees can use agents at work without leaving Google’s enterprise software stack. (cloud.google.com 1) (cloud.google.com 2) For nontechnical users, Google added a no-code Agent Designer that builds trigger-based workflows without programming. For more complex processes, Agent Studio offers a low-code interface for testing multi-step agents in natural language. (blog.google) (cloud.google.com) Google said the platform gives first-class access to more than 200 models through Model Garden, including Gemini 3.1 Pro, Gemini 3.1 Flash Image and Lyria 3. (cloud.google.com) (blog.google) The hardware pitch came alongside the software launch. Google introduced two eighth-generation Tensor Processing Units: TPU 8t for frontier-model training and TPU 8i for large-scale inference and reinforcement learning. (cloud.google.com 1) (cloud.google.com 2) On Google’s TPU page, the company said TPU 8t delivers up to 2.7 times better performance per dollar for large-scale training, while TPU 8i delivers an 80% performance-per-dollar improvement for low-latency inference on large mixture-of-experts models. (cloud.google.com) The launch also clarifies Google’s product reshuffle. Google previously said Agentspace had become part of Gemini Enterprise in October 2025, and the agent creation technology now sits at the center of this broader platform. (blog.google) (cloud.google.com) Google framed the rollout as part of a larger enterprise adoption push, saying 75% of Google Cloud customers now use AI products and its APIs process 16 billion tokens per minute. (www.innobu.com) The immediate test is whether companies use Gemini Enterprise as a central control plane for agents instead of stitching together separate models, tools and workflow software on their own. (cloud.google.com) (blog.google)