RCA and DESSIPI AKT 3

Fashion retailers currently guess which clothes to stock in which stores, often producing too much of the wrong items and too little of what customers actually want. This project builds machine learning and AI models that forecast exactly which clothing types, in what quantities, and at which store locations will sell—replacing intuition with data-driven predictions. The problem is waste. When retailers misjudge demand, unsold clothes end up discounted, donated, or landfilled, while popular items sell out and generate lost revenue. Current planning tools are too crude to handle the complexity of hundreds of store formats, regional preferences, and seasonal shifts. These new models aim to align production directly with consumer demand. If successful, the impact would be felt across supply chains that most people never see: factories producing fewer unwanted garments, warehouses holding less dead stock, and trucks shipping only what stores actually need. For retailers, this means higher profitability and lower markdown costs. For the environment, it means less textile waste and reduced carbon emissions from overproduction. The research is applied rather than fundamental—it takes existing AI techniques and adapts them to a specific commercial problem with measurable outcomes.