Agentic Commerce: A Survey of How AI Agents Are Reshaping Commerce

Recent advances in large language models (LLMs) and agentic AI have catalyzed a paradigm shift from human-driven Traditional Commerce and platform-mediated E-Commerce toward Agentic Commerce (A-Commerce), in which autonomous agents interact, coordinate, and execute commercial activities on behalf of distinct commercial roles. In this survey, we provide the first systematic overview of A-Commerce through a lifecycle-based and agent-role-centric taxonomy that formalizes A-Commerce as a multi-agent setting and analyzes how different agent roles participate across the commerce lifecycle. Our survey then examines how these agentic capabilities are instantiated across real-world application domains, ranging from fully online platforms to hybrid online-to-offline services and physically grounded environments. Further, we review system architectures, protocols, and implementation patterns that enable agentic commerce under real-world constraints, summarize existing benchmarks and evaluation practices, and identify key challenges and promising directions for future research.