Google Cloud: Gemini Enterprise Agent Platform announced at Next '26

At Google Cloud Next ‘26 on April 23, 2026, in Las Vegas, Google announced the Gemini Enterprise Agent Platform — a full-stack system designed to take AI agents from prototype to production in enterprise environments. The announcement covers the entire lifecycle of agent deployment and is organized around four functional areas.

Building: Agent Studio provides a low-code visual interface for teams without deep engineering resources. The upgraded Agent Development Kit (ADK) supports code-first development for teams that prefer programmatic control.

Scaling: A re-engineered Agent Runtime supports long-running agents — those that execute over extended periods rather than completing in a single session. Memory Bank adds persistent context, so agents retain information across sessions without requiring the user to re-establish it each time.

Governing: Agent Identity assigns unique credentials to each agent, enabling access control and audit trails. Agent Registry provides a catalog of deployed agents across an organization. Agent Gateway handles policy enforcement and routing.

Optimizing: Agent Simulation, Agent Evaluation, and Agent Observability tools let teams test, measure, and monitor agent behavior before and after deployment.

Google also announced Gemini Enterprise Projects — a workspace feature that gives Gemini a persistent memory spanning multiple sessions, reducing the overhead of re-establishing context in multi-step work.

Why this matters for product managers

The announcements are significant on two levels. For teams building on Google Cloud, the platform substantially lowers the work required to deploy agents in production — particularly the governance and observability layers, which are currently built from scratch at most organizations. Having agent identity, registry, and policy tooling as managed infrastructure removes a large category of work that product and engineering teams have been treating as a prerequisite to shipping.

For PMs tracking platform strategy, Next ‘26 signals that Google is investing heavily in the agent infrastructure layer as a competitive differentiator against Microsoft and Anthropic. The governance tooling in particular — agent registries and identity systems — addresses compliance questions that enterprise buyers have raised as a blocker to production AI deployments. This type of infrastructure capability tends to move enterprise procurement conversations, which makes it worth tracking even for teams not currently building on GCP.