Customer segmentation through machine learning is a powerful tool that allows companies to create marketing strategies and increase the personal effectiveness of their campaigns. Despite the popularity of methods such as K-means clustering, traditional segmentation approaches often do not reflect the dynamic nature of customer behavior. Transitions between different segments remain outside the scope of static models. This limits the ability to anticipate and adapt to future changes in consumer behavior. This paper explores a hybrid approach that combines the power of clustering and the dynamics of stochastic processes.