Mapping the intellectual structure of artificial intelligence-enabled customer engagement: a bibliometric analysis and framework-based review for future research directions

Artificial intelligence (AI) is reshaping marketing by enabling personalization, automation and interactive experiences. The research on AI-enabled customer engagement (AI–CE) remains fragmented, lacking systematic mapping and an integrated framework. This study aims to examine publication trends, contributors and collaborations; map intellectual and knowledge structures; and develop a framework identifying enablers, challenges and future research opportunities. This study followed a two-stage design combining bibliometric and thematic analysis. Data were retrieved from the Scopus database on February 17, 2025 using a comprehensive set of search terms related to AI and CE, yielding 6,057 records. After applying Preferred Reporting Items for Systematic Reviews and Meta-Analyses-based screening criteria covering publication period (2001–2024), subject areas (computer science, business, management, accounting, decision sciences, social sciences and psychology), document type (journal articles), publication stage (final) and language (English) and eliminating duplicates, 1,241 studies were retained. Bibliometric analysis was conducted using MS Excel, VOSviewer and Biblioshiny (R Studio) to examine publication trends, leading contributors, journals, affiliations and countries. Network mapping through citation, bibliographic coupling and keyword analysis identified intellectual structures. From these, 88 interconnected articles formed the core data set for thematic synthesis, which was carried out using the theory–context–characteristics–methodology (TCCM)-based framework. AI–CE research has expanded from three publications in 2001 to 690 in 2024. Lee emerged as the most prolific author (15 papers), while Li was the most influential (410 citations). IEEE Access, Expert Systems with Applications, Sustainability, and Computers in Human Behavior were top outlets. Leading institutions included Huazhong University of Science and Technology, Griffith University and Northeastern University, with strong USA–China collaborations. Science mapping identified five clusters: (1) AI and digital transformation (business model innovation and value cocreation), (2) AI–human interaction (interface design and transparency), (3) AI-powered service agents (chatbots, trust and brand outcomes), (4) AI service quality (human–AI complementarity and retention) and (5) AI and privacy (trust and ethical concerns). The TCCM synthesis revealed gaps in theory (limited sociopsychological/value-based models), context (dominance of developed economies and select sectors), characteristics (privacy, ethics and algorithmic bias underexplored) and methods (reliance on cross-sectional surveys). The three-layer framework shows that CE acts as both a direct and indirect pathway linking AI determinants to purchase outcomes. This paper presents a substantive conceptual advancement by offering a comprehensive and integrative synthesis of research in this field. It integrates the bibliometric analysis with the unified three-layer framework from foundational knowledge (TCCM) to conceptual pathways mapping AI–CE to purchase intention linking AI determinants, CE and purchase behavior, while systematically identifying theoretical, contextual, characteristic and methodological gaps. This integration links theoretical and empirical insights, transforming TCCM into a decision-oriented tool with sector-specific managerial implications and societal/policy implications. The value of this study lies in guiding future research by highlighting underexplored theories, diverse contexts, ethical and privacy considerations and advanced research design.