You can define, apply, and manage policies that govern agent interactions.
By using the Policies page, you can do the following:
Use Identity and Access Management (IAM) policies to govern agentic communication.
Use Semantic Governance policies to govern traffic between agents and MCP servers and other tools.
Use IAM policies to govern agentic communication
You can create IAM Unified Access Policies (Access policies) that Agent Gateway uses to more securely govern agentic communication between your agents and destination resources, including other agents, MCP servers, and endpoints. Agent Gateway uses Identity-Aware Proxy (IAP) to enforce the policies.
For detailed information about IAM policies for Agent Gateway, see IAM Access policies overview. To create an IAM Access policy, see Create IAM Access policies.
Use semantic governance policies
Semantic governance policies provide a natural language-based security and compliance layer that helps ensure an AI agent's tool invocations align with both user intent and organizational business constraints. While security mechanisms like IAM are static, semantic governance policies handle the non-deterministic nature of large language models (LLMs) by allowing administrators to define security and business rules using natural language constraints (NLC).
To learn more about semantic governance policies, see Semantic governance policies overview. To configure semantic governance policies, see Configure semantic governance policies.
What's next
Codelab: Secure cross-cloud agentic AI applications
Learn how to secure your agentic applications in the Securing Cross-Cloud Agentic AI Applications codelab.