Artificial Intelligence (AI) Assistant in Online Shopping: A Randomized Field Experiment on a Livestream Selling Platform

Reference : Wang, L., Huang, N., He, Y., Liu, D., Guo, X., Sun, Y., & Cheng, G. (2025). Artificial intelligence (AI) assistant in online shopping: A randomized field experiment on a livestream selling platform. Information Systems Research, 36(4), 2358–2374. https://doi.org/10.1287/isre.2023.0103

Key words: livestream selling • artificial intelligence (AI) • AI streaming assistant • human-AI interaction • product return • randomized field experiment.

Summary:

This article studies the impact of an AI-powered streaming assistant in livestream shopping platforms. The research investigates whether an AI assistant can help solve the tension between:

  • streamers’ limited ability to answer all customer questions,
  • and consumers’ need for immediate personalized information

Development: 

The study uses a large-scale randomized field experiment on a major Asian livestream shopping platform to examine how AI assistants affect:

  • purchase behavior,
  • consumer decision-making,
  • and product return rates

Methodology : quantitative study

H1.aThe implementation of an AI streaming assistant during streaming sessions increases the number of purchases in livestream selling.

Hypothesis 1b. The implementation of an AI streaming assistant during streaming sessions decreases the number of purchases in livestream selling.

H2. The implementation of an AI streaming assistant during streaming sessions decreases product return rates in livestream selling

Research Design

The study used:

  • a randomized field experiment
  • on a leading livestream shopping platform in Asia.

Key findings:

AI assistants increase purchases

AI assistants reduce product return rates

AI reduces uncertainty in decision-making

AI works better for uncertain products