Gemini Enterprise Agent Platform enhances agent capabilities and security

Agent Memory Bank allows for low-latency agent personalization by defining structured schemas to extract and maintain critical conversation context. This feature ensures agents retain user preferences, past decisions, and account history across long-running tasks, improving efficiency and user experience.

Agent Runtime enables automation of complex, multi-day agents and reasoning tasks by allowing agents to run continuously for up to 7 days. This facilitates delegation of asynchronous processes, such as sales sequences or multi-stage onboarding, without constant human intervention.

Agent Identity is a new native IAM type built on open standards that enforces least-privilege access to agent permissions. It mitigates token theft, provides non-repudiable auditing of agent actions, and automates identity lifecycle management.

Agent Gateway provides a central control point for securing and governing interactions across an agent ecosystem. It allows for granular access controls via IAM conditions and natural language rules, with integrated protection against prompt injection and data leakage.

Agent Registry serves as a single library for all AI agents, servers, and connections within an organization. This facilitates easy discovery and reuse of agents, improving system organization and enabling administrators to monitor agent sprawl.

The platform now includes continuous monitoring and evaluation of agent performance in production using online monitors that detect degradation and behavioral drift. It also offers comprehensive tracing and real-time dashboards for deep visibility into agent reasoning and execution.