This presentation introduces an approach and application of generative AI in creating dynamic, intelligent honeypots for cybersecurity. Traditional honeypots, while effective, often become obsolete as attackers learn to recognise and avoid them. Leveraging generative AI, we can design honeypots that continuously evolve, presenting new and unpredictable environments that lure and trap sophisticated attackers. This session will showcase a case study of a leading financial institution that implemented AI-powered dynamic honeypots, significantly enhancing their threat detection and response capabilities. We will explore the technical architecture, including the use of advanced generative models to simulate diverse attack surfaces and deceptive data streams. Attendees will gain insights into the practical implementation, challenges faced, and the profound impact on the organisation's overall security posture. This will not only strengthen defences but also provides invaluable intelligence on emerging threat vectors.