The challenge
The global manufacturer struggled with predicting lead times and costs accurately - less than half of non-stockable items arrived on time and errors multiplied across thousands of suppliers.
Late deliveries, frustrated customers
Less than half of non-stockable items arrived on time. Customers were frustrated, inventories grew unnecessarily and costs spiraled.
Manual estimates didn't scale
Estimating supplier lead times and costs manually was slow and often inaccurate. With over 5,000 active suppliers worldwide, errors were unavoidable - and every wrong estimate showed up as a delay, extra cost or lost order.
AI needed business impact, not experiments
The company knew AI could be the answer - but where to start, which use cases mattered and how to make sure the investment paid off? They needed a partner who understands both business and technology.
What we built
A five-week pre-study mapped the bottlenecks, and within six months we put several AI solutions into production - bridging executive vision and technical delivery.
Results
A dramatic shift in the supply chain: more accurate predictions, better inventory turnover and procurement moving from reactive to strategic.
Fewer rush orders, better inventory turnover and happier customers - the machine learning model predicts real arrival times.
Manual estimates were off by EUR 820 per item on average. The AI is off by just EUR 45.
For externally sourced items the improvement reached 23%.
Decisions are based on data, not guesswork. Teams can focus on customers and development while AI handles routine predictions.
