Displaying the main behaviors of customers on a customer journey map (CJM) helps service providers to put themselves in their customers’ shoes. Inspired by the process mining discipline, we address the challenging problem of automatically building CJMs from event logs. In this paper, we introduce the CJMs discovery task and propose a genetic approach to solve it. We explain how our approach differs from traditional process mining techniques and evaluate it with state-of-the-art techniques for summarizing sequences of categorical data.