Planning is a crucial component of processes in many domains, including production lines, general problem-solving, optimization, or many others. However,
planning as a research field often remains difficult to understand and inaccessible to non-experts. To address this issue, the PlanningCopilot is introduced. It combines the PlanPilot planning system with a large language model (LLM) to bridge the gap between complex planning systems and users of varying levels
of expertise. The goal is to allow for an easier usage of PlanPilot and, in doing so, making planning as a whole more accessible to everyone. To evaluate whether this objective was achieved, a user study was conducted and analyzed. The results show that the PlanningCopilot successfully enables users to explore the plan and facet spaces, review the plans, and adapt them to their individual needs. Test participants reported a significantly improved user experience and considerably easier usability compared to PlanPilot. Furthermore, the integration of an LLM also enables new functionalities such as swap detection.
Publication Date: 2026-06-15