As Internal Developer Platforms (IDPs) face the challenge of supporting both traditional and AI development workflows, it is essential to evolve the platform to accommodate the unique requirements of machine learning workloads while maintaining existing functionality. Drawing from practical experience in building enterprise platforms, Max Körbächer will discuss how to evolve an IDP to support AI without disrupting what already works. The session will cover approaches to managing specialized hardware resources, implementing efficient CI/CD for ML models, and enhancing developer experiences for both traditional and AI-driven workflows. Attendees will learn practical strategies for GPU sharing, managing model deployments, and creating self-service capabilities that cater to all developer personas. The session will also provide patterns for platform evolution, real-world examples of AI-ready IDPs, and frameworks for measuring platform effectiveness in the AI age.