As CX organizations move from isolated AI agents to multi-agent environments, the challenge shifts from building successful agents to governing how they act. And, maybe more importantly, how they interact. This interactive workshop with Cyara and Opus Research introduces Opus’ agentic control plane model, outlines its five critical layers, and then dives deeper into evaluation and policy as the foundation for trusted CX automation. Attendees will leave with a practical framework for governing, evaluating, and improving AI agents in real-world CX environments.