Customer segmentation in the automotive industry: from classical clustering to explainable artificial intelligence — A systematic review

. In the automotive sector, customer segmentation has progressed from static, demographically driven models to data-centric and interpretable frameworks supported by artificial intelligence. This systematic review synthesizes 24 peer-reviewed studies published between 2010 and 2025, following the PRISMA 2020 protocol. Three methodological paradigms were identified: (1) classical clustering techniques (RFM, K-Means, Hierarchical analysis); (2) hybrid statistical machine learning approaches (Gaussian Mixture Models, DBSCAN, Logistic Regression); and (3) AI-driven segmentation using explainable techniques like SHAP and LIME. This study shows a progressive shift from descriptive segmentation to predictive and interpretable models. While Evaluation metrics such as the Silhouette and Davies-Bouldin are fundamental to assess cluster compactness and separation, they do not explain the model interpretability. In Morocco, segmentation remains static but it is slowly evolving towards intelligent strategic marketing enriched by data and assisted by AI. This study contributes to the managerial and methodological literature by highlighting the key trends in segmentation practices in emerging markets.