General practice is facing a global crisis due to a shortage of GPs and increasing workload stress, leading to burnout and fewer new doctors entering the field. This situation is particularly severe in countries like Switzerland, where the number of active GPs has been steadily declining. The stress of administrative duties and the demands of patient care worsen the problem, pushing many GPs to reduce their workload or leave the profession entirely.

This study evaluates the current capabilities of LLMs in assisting GPs by focusing on both administrative and clinical tasks. Through web research and interviews with healthcare professionals in Switzerland, the research identifies key pain points in general practice. The study then tests the effectiveness of various LLMs in addressing these issues, focusing on tasks like report writing, diagnosis assistance, and proposes a tailored AI-powered assistant for GPs.

The evaluation reveals that LLMs hold significant promise in reducing the administrative burden on GPs, allowing them to focus more on patient care. The proposed AI assistant, designed to automate routine tasks and provide quick access to medical knowledge, shows potential for improving efficiency and job satisfaction among GPs. Initial tests demonstrate that LLMs can accurately handle many routine tasks. However, further refinement is needed to ensure that the AI can fully integrate with existing healthcare systems and handle the complexities of individual patient cases. The thesis concludes by emphasizing the need for ongoing development, including clinical trials and regulatory alignment, to ensure that the AI assistant can be effectively deployed in real-world general practice, contributing to a more sustainable and less stressful healthcare environment.