AI-Integrated Buy Now, Pay Later (BNPL) Schemes in E-Commerce: A Dual-Method Systematic Review and Biblio-metric Analysis of Stock Market Implications (1991–2025)

The convergence of artificial intelligence (AI) and Buy Now, Pay Later (BNPL) services constitutes a structural transformation in electronic commerce, yet its implications for capital market valuations remain theoretically underdeveloped. This study deploys a dual-method approach, integrating PRISMA-guided systematic literature review with VOSviewer-driven bibliometric analysis, to critically examine 187 publications indexed in Web of Science (WoS) and Scopus between 1991 and 2025. Drawing upon a critical synthesis of the 50 most impactful studies from 2020 to 2025, we identify three mechanism-driven knowledge clusters: (1) AI-enabled credit risk architectures within BNPL ecosystems, (2) fintech disruption signals and equity market reactions, and (3) regulatory technology (RegTech) feedback loops on BNPL provider valuations. Our gap analysis reveals that existing scholarship overwhelmingly privileges consumer behavioral outcomes while neglecting the transmission mechanisms through which AI-integrated BNPL innovations propagate into stock market pricing efficiency. We advance five testable hypotheses linking algorithmic underwriting sophistication, default prediction accuracy, and merchant network externalities to abnormal stock returns among publicly listed BNPL providers. The timeline analysis identifies 2019 as the inflection point where machine learning integration shifted BNPL from a payment's convenience to a data-driven financial instrument. We propose a research agenda targeting the identified gaps and predict five emerging trends including federated learning for cross-border BNPL risk assessment and generative AI-driven dynamic credit limit optimization. This review contributes to the theoretical and applied electronic commerce literature by articulating a causal framework connecting AI capabilities, BNPL market structure, and investor valuation responses.