This semester, the Chair of Information Systems and Supply Chain Management, headed by Prof. Bernd Hellingrath, is offering the project seminar “Agentic Purchase Recommendations for Inventory Management Optimization” for master’s students. In collaboration with the Münster-based company pacemaker.ai, the students are investigating how agentic AI systems can support supply chain planners in their day-to-day work, particularly with replenishment decisions and demand forecasting.

Planners face complex decisions every day: volatile demand, supply bottlenecks, and a large number of materials, suppliers, and constraints all have to be kept in view at once. Agentic AI opens up new possibilities here. Unlike traditional forecasting models, AI agents can independently gather information, detect anomalies, explain their recommendations, and interact with users in natural language. The central question of the seminar is therefore: How should such a digital “sidekick” be designed so that it genuinely relieves planners of work while keeping humans in charge of the decisions?

Working in small teams, the students will analyze existing planning processes and planners’ requirements and review the current state of research on agentic AI and human-AI collaboration. Building on this, they will develop their own concepts and prototypes, such as an agent that explains forecast deviations in an understandable way, an assistant that prepares order proposals along with their rationale, or a system that proactively flags critical shortages. The teams will be supported by experts from pacemaker.ai, who will provide insights into real-world planning processes, data, and practical requirements.

pacemaker.ai is headquartered in Münster, is part of the thyssenkrupp group, and develops AI software for industrial supply chains, with solutions for demand forecasting and replenishment.