There is a constant conflict in global corporations between satisfying the business needs and complying with the travel policy to avoid paying over-priced fares. This article reviews the nature of business travel, discusses its main characteristics, and the complexity for management in defining the policy. A travel policy must be feasible to comply with and contribute to optimizing the budget. Each company requires a different policy according to business needs; a manufacturing company with planned trips requires a different policy than a consulting firm with unpredictable trips. We adapted the machine-learning algorithm of regressive decision trees to include the characteristics of business travel policy; our algorithm self-adjusts to each company's data. We test our results using travel data from a company and simulating different scenarios. Our findings support management with a quantitative method to determine the optimal travel policy.
JEL Classification: M10 , M19 , C610 , C630