Reliability is the core challenge for AI agents. AI agents are systems whose components, such as LLMs, can be unreliable. This presentation will introduce CIOs to a set of principles and best practices to design AI agents with reliability as a key principle.