At Google Cloud Next, Sundar Pichai unveiled the Gemini Enterprise Agent Platform, a system designed for constructing and operating AI agents in production. This move directly counters Amazon's Bedrock AgentCore and Microsoft Foundry. The notable architectural decision is its primary user: Google built this platform for IT and technical teams, not general business units.

This focus reflects a pragmatic engineering assessment. Current agent technology shows the most reliable performance on technical workloads, such as code generation, where tasks have clearer boundaries and success criteria. Deploying novel AI systems at enterprise scale also introduces significant security and governance complexity, which technical teams are best equipped to manage.

Business users are instead steered toward the separate Gemini Enterprise app for lighter, workflow-specific tasks. The division creates a clear separation of concerns: engineering builds and maintains the core agent infrastructure, while other departments consume approved capabilities or assemble simpler automations.

Under the hood, the platform offers model flexibility, integrating Google's Gemini LLM and Nano Banana 2 alongside third-party options like Anthropic's Claude suite—Opus, Sonnet, and Haiku. This multi-model support allows engineers to make cost-versus-performance trade-offs based on specific agent functions, a necessary consideration for scalable system design.

Source: TechCrunch