Enabling Human-AI Collaboration in a Modular Planning Environment

The vision of fully automated factory planning remains an elusive goal. The complexity and context-dependency of tasks in factory panning still demand human expertise. This paper introduces the concept of a Factory Planning Copilot, which shifts the focus from automation to intelligent assistance. Inspired by systems like Microsoft Copilot, the approach envisions a virtual environment where planners interact with digital factory models through natural language, supported by large language models. Rather than replacing human judgment, the Copilot augments it by interpreting intentions, suggesting improvements, and enabling intuitive access to complex planning data. By embedding domain knowledge through modular expert systems and fostering an open ecosystem, the Copilot can accelerate planning processes, enhance decision quality, and improve compliance with internal and external requirements. Copilot for Factory Planning aims to make digital factory planning more accessible, adaptive, and reliable to empower planners through seamless human-AI collaboration.