Abstract
The rapid growth of digital commerce has made fraud across social media platforms and online marketplaces increasingly complex, posing significant challenges for organizations and policymakers alike. This study identifies the technological and organizational design requirements for developing an AI-enabled fraud detection framework suited to Kazakhstan's digital commerce environment. A configurative literature review was conducted following the PRISMA 2020 guidelines, drawing on the Scopus and Web of Science databases. From an initial sample of 376 publications, more than 20 peer-reviewed studies published between 2020 and 2026 were retained after screening and eligibility assessment. A bibliometric analysis using VOSviewer further mapped dominant research themes and their conceptual relationships. The findings show that existing research remains predominantly technology-driven, while organizational governance, implementation, and project management have received comparatively little attention. Based on this synthesis, the review identifies five major thematic areas. It proposes evidence-based design requirements, including multi-source data integration, hybrid AI architectures, governance and compliance mechanisms, human-in-the-loop decision-making, and continuous monitoring – offering a conceptual foundation for an AI-enabled fraud detection framework for Kazakhstan's digital commerce sector.