This research employs a serial mediation model to explore how different levels of human-robot collaboration, apology styles, and emotional responses affect customer intentions after a service failure, based on a scenario-based experiment with 311 participants, analyzed using MANCOVA and PROCESS Macro Model 6. Our findings reveal that human-robot collaboration configurations where robots play a significant role, either augmenting or replacing humans, are more effective. Economic apologies are more impactful when the robot leads the recovery, while social apologies work best when human staff are involved. Comfort emotions and robot continuance usage sequentially mediate the relationship between human-robot collaboration and behavioral intentions. This is the first paper to integrate frontline technology with traditional recovery methods, highlighting the effectiveness of human-robot collaboration in enhancing customer retention. Practically, this research provides essential guidance for robot and AI designs in services, enabling service managers to effectively manage human-robot task allocation and customer loyalty in a robot-mediated service recovery.

中文翻译：

---

服务恢复中的人机协作：检查道歉风格、安慰情绪和客户保留率


本研究采用串行中介模型，基于对 311 名参与者的基于场景的实验，使用 MANCOVA 和 PROCESS 宏模型 6 进行分析，探讨不同级别的人机协作、道歉风格和情绪反应如何影响服务失败后的客户意图。我们的研究结果表明，机器人发挥重要作用的人机协作配置，无论是增强还是取代人类，都更有效。当机器人领导恢复时，经济道歉的影响力更大，而当人类员工参与时，社会道歉效果最好。安慰情绪和机器人连续使用依次中介人机协作与行为意图之间的关系。这是第一篇将一线技术与传统恢复方法相结合的论文，强调了人机协作在提高客户保留率方面的有效性。实际上，这项研究为服务中的机器人和 AI 设计提供了必要的指导，使服务管理者能够在机器人介导的服务恢复中有效地管理人机任务分配和客户忠诚度。