The clinical diagnosis of sepsis is challenging not only in outpatients but also critically ill patients in the intensive care unit. In the latter setting, sepsis has been recognized as suitable use case for prediction modelling, but most prediction models never arrive in clinical practice. In my presentation, I will provide an overview of sepsis as a medical issue including its treatment and will present current measures for improvement of treatment quality and lowering mortality in Germany. I will then focus on existing prediction models and the gaps arising in their validation which I intend to tackle with the BMFTR-funded project “Time, Transparency, Transportation and Transfer for bridging the AI-prediction model validation gap” (4T4AI).