Advanced scientific user facilities, including self-driving laboratories, are revolutionizingscientific discovery by automating repetitive tasks and enabling rapidexperimentation. However, these facilities must continuously evolve to supportnew experimental workflows, adapt to diverse user projects, and meet growing demandsfor evermore sophisticated instrumentation. This continuous developmentintroduces significant operational complexity, necessitating a focus on usability,reproducibility, and intuitive human-instrument interaction. In this work, we explorethe integration of agentic AI, powered by Large Language Models (LLMs),as a transformative tool to achieve this goal. We present our approach to developinga pipeline for operating a robotic station dedicated to the design of novelmaterials. Specifically, we evaluate the potential of various LLMs as trainablescientific assistants for orchestrating complex, multi-task workflows, optimizingtheir performance through human input and iterative learning. We demonstratethe ability of AI agents to bridge the gap between advanced automation and userfriendlyoperation, paving the way for more adaptable and intelligent scientificfacilities.