Agentic AI is now executing live enterprise workflows at scale — approving transactions, modifying records, chaining decisions across systems. Most boards and CEOs have yet to answer the most consequential governance question: when an AI agent causes harm, who owns it?
