Dan Finan sits down with Tristan Watkins, Head of Service Architecture at Advania, to break down AI governance beyond the buzzwords. Tristan brings 16 years of experience in security and compliance to tackle the real challenges organisations face when implementing AI safely. From the EU AI Act timeline to managing shadow AI usage, this conversation cuts through the complexity to focus on practical risk management strategies.

Tristan explains why traditional IT problems like data permissions and overshared information remain the most critical AI governance challenges. He walks through the specific risks of services like Deep Seek, including advertising technology exploitation and data correlation concerns that most organisations overlook. The discussion covers everything from private ChatGPT solutions to AI agent safety controls, prompt injection attacks, and why bias evaluation frameworks often prove impractical for most projects.

This episode delivers actionable insights for IT leaders navigating AI implementation while maintaining security and compliance standards.