AWS AI and Data Conference 2026 - Building Safe AI Agents | AWS Events
Learn how to deploy AI agents with controls that restrict what they can say and do. This session covers Amazon Bedrock Guardrails that filter agent content (input/output) and Amazon Bedrock AgentCore Policy that enforces rules on agent actions. Attendees will learn how to: - Choose between guardrails (controlling input/output) and policy (controlling actions) - Define policies that enforce rules on agent actions - How AWS uses automated reasoning to validate controls.