<p>This paper addresses the scheduling problem of filling customer pill orders on a set of automated dispensing machines. Each order consists of a 1-month supply of multiple pill types with varying quantities that can be assigned to and processed by the machines. The client company had previously used a first-come-first-served (FCFS) rule to assign orders to machines, resulting in frequent stoppages to refill canisters. We introduced a smart scheduling program with the objective of reducing machine downtime by having as many pill canisters as possible become nearly empty at the same time. In comparison with the FCFS approach, our smart scheduling program reduces the number of machine stoppages (i.e., downtime) by an average of 83% and increases daily effective capacity by approximately 60%, without adding labor or equipment. This work brings new perspectives and tools to the management of differentiated customers in complicated order-filling environments such as those seen in online pharmacies and health supplement companies.</p>