Artificial intelligence (AI) is undergoing a fundamental shift. Systems that once served as assistive tools (e.g., predictive models and generative AI) are now evolving into agentic systems capable of planning and executing complex, multistep tasks with minimal human input. As AI systems’ ability to plan and execute tasks autonomously accelerates, so too does the potential speed and scale of cyber-attacks. Compounding this risk, traditional cyber-defences were designed to react to human-driven threats rather than automated, machine-speed attacks. As such, organisations may be increasingly vulnerable to losses arising from cyber incidents unless their security frameworks evolve to keep pace.