Philipp Herzig explains SAP's resilience through technological transitions, from mainframe to AI, by evolving from custom to standard software. As CTO, his top priorities include AI strategy and re-engineering systems, with all employees using agentic coding. The discussion highlights the challenges of enterprise AI adoption, including data fragmentation, security risks, and the need for verifiability. Herzig contrasts LLMs with classical ML for predictive tasks, introducing SAP's RPT1 transformer for structured data. He predicts AI will uplevel work by automating mundane tasks and shift SAP's pricing from seat-based to outcome-based models. The conversation also touches on the potential of quantum computing for optimization problems in logistics and supply chain, emphasizing that complex enterprise issues still challenge current algorithms.