Abstract
Artificial intelligence (AI) has become a core enabler of customer data analysis in e-commerce, yet existing research remains fragmented across applications, methodologies, and contexts. This study presents a systematic literature review (SLR) of peer-reviewed research published between 2018 and 2025 on AI-based customer data analysis in e-commerce, conducted in accordance with the PRISMA 2020 framework. Following a structured search across Scopus, Web of Science, and Google Scholar and the application of pre-defined inclusion and exclusion criteria, fifteen studies were included in the final synthesis. Thematic analysis identified five interconnected domains of AI application: recommendation systems, sentiment analysis, customer behavior prediction, customer service automation, and privacy and security. Findings indicate a consistent shift from classical statistical methods toward deep learning and transformer-based architectures across all five domains, alongside measurable improvements in personalization, response efficiency, and customer satisfaction. However, the review also identifies significant technical, ethical, and methodological gaps, including limited multilingual and cross-platform support, inconsistent quantitative reporting, and underdeveloped privacy-preserving frameworks. This study contributes an integrated thematic framework and a research agenda comprising seven priority directions, offering practical guidance for e-commerce practitioners and a foundation for future empirical and cross-cultural research.