Foundation models are the basis of the new AI paradigm, generative AI, which has been on everyone’s lips since „ChatGPT“. Foundation models are pre-trained large AI models that can be easily used, e.g. to support text analysis or text generation. The trick is to adapt these AI models to your own needs, i.e. your own language, either through prompt tuning or through fine-tuning. Technologically, open source software, on-premise operation and a trustworthy AI „à la EU AI Act „represent important skills. How this is possible in practice and concrete examples from various industries will be shown in the lecture.