EDB's Chief Engineering Officer, Max Romanenko, emphasizes that AI agents, unlike human users or applications, require a new set of access, decision-making, and governance requirements from day one due to their 24/7/365 operation and rapid workforce expansion. He notes that agent identity and access management must be context-specific, tied to individual tasks, and enforced at the data layer using primitives like Postgres's roles and row-level security.
Enterprise AI agent deployment is bottlenecked by data governance and access control, requiring new infrastructure solutions beyond traditional IAM.