Generative AI Systems play an increasingly important role in Software Engineering. In order to systematically analyze (1) how to use this technology in specific contexts, and (2) to come to conclusions, whether massive investments into application of these technologies should be made, we try so setup several real-world experiments and case studies. We therefore aim at laying the basis for informed decisions on the systematic application of LLMs technologies for Software Engineering for specific application scenarios from an efficiency perspective. The focus is on supporting migration of systems, on software testing, and on automatic design refactoring.