Securing the Agentic Enterprise | Interview with Charlie Lewis

Host Dejan Kosutic interviews McKinsey & Co. partner Charlie Lewis about how agentic AI will reshape cybersecurity by reinventing existing categories — especially identity, detection, operations, and vulnerability management — rather than creating entirely new ones. Lewis argues every company will manage a massive digital workforce of dynamically created, short-lived agents with constantly changing permissions, pushing traditional IAM beyond its design and requiring workforce-plus-agent identity, stronger data governance, and an “agent registry.” They discuss the need to strengthen foundational controls like IT asset management, map business processes to prioritize attack paths, and shift security from periodic detection to continuous runtime governance with real-time policy enforcement. Lewis predicts more machine-operated, human-supervised SOC work, highlights in-demand skills, and advises CISOs to embed with the business, fix processes before automating, and pilot agentic use cases in security. (00:00) - Interview with Charlie Lewis (01:11) - Why Cyber Reinvention (02:46) - Nonhuman Identity Frontier (06:00) - Vendor Readiness Today (08:20) - Org Change Data Governance (11:40) - Asset Management Agent Registry (14:38) - Mythos Runtime Governance (17:53) - Future SOC Human Role (27:37) - Consulting Role Evolution (29:56) - Vendor Recommendations Summary (33:44) - Buying Beyond The CISO (40:07) - Security as Business Enabler (43:27) - Top CISO Recommendations