Agentic Commerce: A Systematic Review of AI-Driven Autonomous Shopping and Emerging Transaction Protocols

Abstract

This study presents a systematic literature review of agentic commerce, with particular emphasis on autonomous shopping systems, machine-to-machine transaction protocols, and the governance conditions necessary for their large-scale adoption. Rapid advances in large language models, generative artificial intelligence, and autonomous agent architectures are transforming digital commerce by enabling intelligent agents to search for and assess products, compare suppliers, negotiate transaction terms, execute purchases, coordinate logistics, and manage post-purchase activities with limited human involvement. Despite increasing academic and industry interest, the relevant literature remains fragmented across e-commerce, information systems, artificial intelligence, cybersecurity, digital governance, and consumer behavior research. Following the PRISMA 2020 methodology, this review examines studies published between 2020 and 2026 and retrieved from Scopus, Web of Science, IEEE Xplore, ScienceDirect, SpringerLink, and the ACM Digital Library, supplemented by searches of Google Scholar and arXiv. Following identification, deduplication, screening, eligibility assessment, and quality appraisal, 42 studies were included in the qualitative synthesis. The evidence was analyzed using the AI-Commerce Integration Assessment Framework, which enabled systematic interpretation of the literature across the dimensions of readiness, feasibility, performance, and maturity. The findings indicate a clear transition from AI-assisted product discovery and recommendation systems toward increasingly autonomous purchasing environments in which AI agents operate as active economic actors. However, broader adoption remains limited by concerns related to trust, privacy, transparency, accountability, cybersecurity, user control, and legal responsibility. The review also identifies emerging infrastructures for digital identity, authentication, interoperability, automated negotiation, smart contracts, payment settlement, and dispute resolution. Nevertheless, these mechanisms remain weakly standardized and insufficiently interoperable across platforms and jurisdictions. Overall, agentic commerce is conceptualized as a transition from human-centered digital transactions toward AI-mediated economic ecosystems. Its sustainable development will depend on the coordinated evolution of technical standards, governance structures, explainable AI, consumer protection mechanisms, accountability frameworks, and secure transaction infrastructures. This review provides an integrated synthesis of the field and identifies priorities for future empirical research on real-world implementation, interoperability, regulation, market effects, and human–AI collaboration.