The collaboration between humans and Artificial Intelligence (AI) agents is founded on distinct strengths, resulting in what is known as ‘hybrid intelligence.’ In the foreseeable future, AI agents and generative AI systems, such as ChatGPT/GPT-4, will become widely accessible as powerful assistance tools. These systems have diverse applications across various occupational domains and responsibilities. A central question of our research emerges: What interactions between users and systems, as well as among different systems, are necessary to technically implement hybrid AI processes and scalable mentoring applications? In the development of our AI-supported mentoring tool, the Mentoring Workbench (MWB), we drew upon research from previous research studies, user studies and testbed evaluations. Our findings highlight essential user-system and inter-system interactions for effective technical implementation of scalable mentoring in university teaching. These interactions include user-friendliness, knowledge management, feedback loops, privacy, data integration and -management, and scalability. We hope that our research serves as a robust foundation for future investigations in this field.