Demand forecasting, inventory optimization and agentic AI – clear, quantitative and practical.
Dr. Jan Fränkle
I am a physicist and hold a doctorate in economics, and I have been developing AI solutions for the supply chain for more than 20 years – today increasingly with agentic AI. What drives me is the question of how to make good decisions when the future is uncertain. As a trainer I do not want to talk about AI. I want to show what it actually does in the everyday work of planning and replenishment, and where its limits lie. What pleases me most is the moment when participants recognise the potential in their own company.
How the forecast, the safety stock and the replenishment decision fit together, and where AI actually improves classical replenishment. With worked examples and live demos in which participants replenish for themselves.
For: Replenishment planners, material planners, supply chain leads
A forecast is a distribution, not a single value. We look at how to measure forecast quality properly, and which metrics lead you astray.
For: Demand planners, S&OP leads, controlling
AI agents can already take on parts of planning and procurement. We look at where they fail, and how to bound a first deployment sensibly.
For: Supply chain leaders, heads of digitalisation
A sober overview of what machine learning can do and what it cannot. This includes how to judge AI promises from vendors.
For: Decision-makers without a technical background
I also offer every topic as in-house training, adapted to the company's industry and data.
In my seminars, participants calculate for themselves instead of reading slides. I work with interactive demos in the browser. You change a safety stock, for example, and immediately see what happens to availability and inventory. The same logic applies to the forecast: a distribution can be shifted, and the consequences for the replenishment decision become visible at once.
In between there are small competitions. Who replenishes a simulated season best? We look at the result together. The point is to retrace one's own decisions.
I make typical errors of reasoning tangible before I explain them. A higher safety factor does not bring a linear gain in availability. MAPE can make a forecast look good that is of little use for replenishment.
Every seminar ends with the question of what can be put into practice in one's own company. There are also detailed handouts to read afterwards.
