Artificial intelligence and behavioural customer segmentation in digital marketing: A systematic literature review of techniques, contexts, and effectiveness

Abstract

As artificial intelligence reshapes digital marketing, the ability to segment customers based on behavioural data has emerged as a critical driver of personalisation, targeting precision, and marketing performance. Yet despite growing scholarly interest, the mechanisms through which AI-powered behavioural segmentation translates into concrete marketing effectiveness remain theoretically fragmented and lack empirical consolidation. This systematic literature review addresses this gap by synthesising evidence on how AI-based behavioural segmentation is conducted, evaluated, and integrated into digital marketing systems. Following PRISMA 2020 guidelines, a structured multi-database search across ScienceDirect, Scopus, and Web of Science identified 117 records, of which 33 peer-reviewed articles published between 2015 and 2025 met the predefined inclusion criteria. To be included, studies had to apply AI or machine-learning techniques to behavioural customer segmentation in digital marketing and be peer-reviewed journal articles or reviews written in English. The reviewed studies employ a range of AI approaches, including clustering algorithms, supervised classifiers, deep learning architectures, and hybrid models, applied to transactional, clickstream, and behavioural customer data. Quantitative designs predominate across the corpus, with a growing shift toward hybrid and explainability-oriented frameworks, driven by the recognised limitations of purely algorithmic approaches in delivering managerially actionable insights. The findings reveal that AI-powered behavioural segmentation can enhance marketing effectiveness, but this relationship is not automatic: it is contingent on how segmentation outputs are operationalised within downstream marketing actions rather than on algorithmic performance alone. Three theoretical contributions are advanced: reconceptualising AI-based segmentation as a managerial bridge between data analytics and marketing decision-making; challenging the implicit equivalence between segmentation quality and marketing effectiveness; and repositioning explainable AI (XAI) as a theoretical requirement rather than a technical feature. Persistent challenges related to dataset generalisability, cultural transferability, ethical governance, and predictive integration are identified, and a structured agenda for future research is proposed.