The rapid evolution of the automotive sector, especially in Advanced Driver Assistance Systems (ADAS), demands increasingly agile, scalable and reliable validation and testing pipelines. In this context, Large Language Models (LLMs) offer a real opportunity to improve industrial workflows by automating complex and technical tasks. This thesis, developed in collaboration with the Nardò Technical Center (Porsche Engineering Group GmbH), creates the foundation for a system that automatically generates structured Test Catalogues in JSON format. The goal is to transform functional requirements into technical test catalogue files. This process requires not only semantic reasoning, but also the ability to manage and interpret large amounts of technical data.