AI agents can revolutionize process optimization by learning, adapting, and executing workflows autonomously in real-time. However, to fully harness their potential, businesses must redesign their processes around these agents, rather than attempting to integrate them into existing legacy systems. This requires a fundamental shift towards an "agent-first" approach, where AI systems operate processes and humans focus on setting goals and defining policy constraints. By doing so, companies can unlock the full potential of AI agents and achieve dynamic optimization. The traditional method of bolting AI onto fragmented legacy workflows using conventional optimization techniques is no longer sufficient. Embracing an agent-first mindset enables enterprises to leverage AI's capabilities to streamline and enhance their operations, making them more efficient and effective1. This matters to practitioners as it necessitates a rethink of their approach to process optimization, prioritizing the integration of AI agents as central components rather than add-ons.