Birmingham City University and Panic Deliveries Limited (T/A PDS Direct Limited) KTP 24_25 R4

A fragile goods transport company is combining its warehouse and delivery data with artificial intelligence to pack vehicles more efficiently and plan smarter delivery routes. The problem is straightforward but costly: when you ship delicate items like glassware, electronics, or medical equipment, you cannot simply stack boxes to the ceiling. Every empty space in a lorry or missed opportunity to combine two deliveries into one trip adds fuel, labour, and vehicle costs. Currently, most small-to-medium hauliers rely on human planners or basic software that cannot juggle the competing demands of fragility, weight distribution, delivery windows, and fuel economy all at once. This project builds a single AI system that learns from the company’s own operational data—what breaks, what routes work, which vehicle configurations save time—and then recommends optimised loads and routes in real time. If it succeeds, the company expects lower per-delivery costs and fewer damaged goods. For the wider logistics sector, the approach could demonstrate how small operators, not just Amazon-sized firms, can use machine learning to cut waste without expensive custom software. The result would be quieter, cheaper, and more reliable deliveries for anyone who orders something breakable online.