Optimizing AI strategies in e-commerce customer service: An agent-based simulation

Authors

• Yali Zhang, Northwestern Polytechnical University
• Zhenbin Ding, Northwestern Polytechnical University
• Jun Sun, The University of Texas Rio Grande ValleyFollow
• Xingfang Zhao, Northwestern Polytechnical University
• Xu Hu, North China Electric Power University
• Zhaojun Yang, Xidian University

Document Type

Article

Publication Date

9-4-2025

Abstract

The growing integration of chatbots in e-commerce customer service presents opportunities and challenges for online retailers in shaping effective artificial intelligence (AI) strategies. This study evaluates human-only, AI-only, and human–AI collaboration strategies using an agent-based simulation model across varying levels of task complexity, service volume, and product margin. Results show that the AI-only strategy excels in low-volume, simple tasks due to its cost-effectiveness, while the human–AI collaboration strategy proves superior in managing high-volume or complex inquiries by scaling human involvement to meet demand. For high-margin products, this collaborative approach delivers the best service, whereas the AI-only strategy is optimal for low-margin items. Enhancing chatbots’ anthropomorphic qualities could further improve service performance, but only if technological advancements are sufficient. The findings provide actionable insights for optimizing AI deployment and fostering adaptive customer service.

Comments

Reprints and permissions

https://rdcu.be/e3UDT

Recommended Citation

Zhang, Y. et al. (2025) “Optimizing AI strategies in e-commerce customer service: An agent-based simulation,” Electronic Markets, 35(1), p. 77. https://doi.org/10.1007/s12525-025-00821-8.

Publication Title

Electronic Markets

DOI

10.1007/s12525-025-00821-8