Governing the agents: Leadership dilemmas in deploying agentic AI in a cyber defence project team

This teaching case follows Maya Mercer, Head of Cyber Defence at ‘Meridian Water & Power’, a United Kingdom critical national infrastructure (CNI) operator, as the security team introduces agentic artificial intelligence (AI) into its security operations centre (SOC). Over roughly 2 years, Mercer’s team moves from a fragmented, alert-saturated environment to one in which three custom multi-agent systems support incident analysis, detection engineering and threat intelligence under human-in-the-loop oversight. The case culminates in a decision: with strong early returns and growing pressure to pursue ‘frontier’ status, should Mercer approve the next phase – granting agents the autonomy to take limited containment actions without waiting for a human – or continue with decision-making by human analysts while the foundation matures? The case raises issues around the use of autonomous AI agents in security management, workforce anxiety and role evolution, and the governance of agentic AI in regulated environments. It is suitable for courses on digital and AI transformation, AI strategy, cybersecurity leadership, and technology and operations management at advanced undergraduate, postgraduate and executive levels.