In the mail service industry worldwide, missed deliveries are common, especially in dense urban areas where residents aren’t usually home when their parcel arrives. PostNord, which operates in the Nordic countries of Sweden and Denmark, was no stranger to this problem. The company knew missed parcels was costing both money and customer satisfaction. With reams of historical data at their fingertips, IBM’s Data Science and AI Elite team put Watson Studio to the test. During a recent 12 week engagement, IBM data scientists developed a machine learning pipeline that revealed where half of those missed deliveries could be avoided -- and by delivering that insight straight to the sorting phase, PostNord could take steps to adjust route planning and mitigate risks in advance.