Organizations deploying agentic AI systems face a governance gap: existing software inventory frameworks were not designed to handle the dynamic, multi-component nature of autonomous AI agents, spanning third-party foundation models, fine-tuned layers, plugins, and runtime tool permissions, creating supply chain blind spots with no standardized documentation practice to close them. AI Bill of Materials (AI BOM) frameworks are emerging as the primary control response, designed to track model provenance, training data lineage, and runtime action scope in ways traditional SBOMs cannot. Without AI BOMs in place, security teams cannot determine whether a compromised upstream model or plugin has reached production, cannot scope incidents involving autonomous agent actions, and cannot demonstrate compliance with emerging AI governance obligations.