The rapid rise of AI agents presents urgent challenges in authentication, authorization, and identity management. Current agent-centric protocols (like MCP) highlight the demand for clarified best practices in authentication and authorization. Looking ahead, ambitions for highly autonomous agents raise complex long-term questions regarding scalable access control, agent-centric identities, AI workload differentiation, and delegated authority. This whitepaper is for stakeholders at the intersection of AI agents and access management. It outlines the resources already available for securing today’s agents and presents a strategic agenda to address the foundational authentication, authorization, and identity problems pivotal for tomorrow’s widespread autonomous systems.