期刊
JOURNAL OF MARKETING
卷 85, 期 2, 页码 14-32出版社
SAGE PUBLICATIONS INC
DOI: 10.1177/0022242920956676
关键词
AI coach; artificial intelligence; aversion; information overload; new technology; sales training; sales force management
类别
资金
- National Natural Science Foundation of China [71925003]
- Shanghai Philosophy and Social Science Plan [2017BGL019]
- National Science Foundation of China [91746302]
The results of the experiments show that bottom-ranked agents face information overload when receiving AI coaching, while top-ranked agents prefer human coaching. By redesigning the AI coach and combining AI with human coaching, performance of agents can be significantly improved, solving the challenges faced by both bottom- and top-ranked agents. These findings provide valuable insights for researchers and managers on the use of AI coaches.
Firms are exploiting artificial intelligence (AI) coaches to provide training to sales agents and improve their job skills. The authors present several caveats associated with such practices based on a series of randomized field experiments. Experiment 1 shows that the incremental benefit of the AI coach over human managers is heterogeneous across agents in an inverted-U shape: whereas middle-ranked agents improve their performance by the largest amount, both bottom- and top-ranked agents show limited incremental gains. This pattern is driven by a learning-based mechanism in which bottom-ranked agents encounter the most severe information overload problem with the AI versus human coach, while top-ranked agents hold the strongest aversion to the AI relative to a human coach. To alleviate the challenge faced by bottom-ranked agents, Experiment 2 redesigns the AI coach by restricting the training feedback level and shows a significant improvement in agent performance. Experiment 3 reveals that the AI-human coach assemblage outperforms either the AI or human coach alone. This assemblage can harness the hard data skills of the AI coach and soft interpersonal skills of human managers, solving both problems faced by bottom- and top-ranked agents. These findings offer novel insights into AI coaches for researchers and managers alike.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据