Chatbot symbolic recovery and customer forgiveness: a moderated mediation model

Abstract

Purpose

Artificial intelligence-based chatbots are frequently used to handle customer complaints in the hospitality and tourism sectors; however, little is known about their recovery strategies. Further, the widespread usage of chatbots is anticipated to affect customers' favorable responses. Therefore, this study aims to examine how chatbots’ symbolic recovery influences customer forgiveness through customer empathy and explore the moderating effect of time pressure on it. Moreover, it investigates the effect of customer forgiveness on customer reconciliation and customer continuous trust.

Design/methodology/approach

Structural equation modeling was used to analyze data collected from 994 customers who have experienced chatbot recovery in tourism and hospitality during the past four months.

Findings

The results show that chatbots’ symbolic recovery stimulates customer forgiveness, which subsequently positively affects customer reconciliation and customer continuous trust. Moreover, customer empathy partially mediates the effect of chatbots’ symbolic recovery on customer forgiveness, and time pressure plays a moderating role in the relationship between chatbots’ symbolic recovery and customer forgiveness.

Practical implications

The results offer highly persuasive insights that may be used to promote chatbots’ symbolic recovery in tourism organizations. The effectiveness of chatbots’ symbolic recovery in achieving customer forgiveness will motivate tourism organizations to use chatbots efficiently in service recovery.

Originality/value

This study extends the theoretical scope of chatbot research by investigating the symbolic recovery capabilities of chatbots. Moreover, it expands the application of SOR theory in the context of chatbot service recovery and reveals the underlying mechanism behind the impact of chatbots’ symbolic recovery on customer forgiveness, thus building and testing an integrative model of chatbot service recovery.

研究目的

系统评估阻碍因素作为技术接受模型（TAM）的先驱方面存在一定的空白。本研究调查了三个阻碍因素, 即不适感、不安全感和风险。此外, 本研究提出了调节变量 - 个人能力（PC）, 并测试其对感知有用性（PU）、感知易用性（PEU）和行为意图（BI）之间关系的影响。

研究方法

使用量化数据分析验证了通过Smart PLS4使用的调整模型。对327名有效受访者的数据进行了分析。

研究发现

不适感是影响PU和PEU的显著先驱因素。不安全感和风险分别是PEU和PU的抑制因素。本研究在稀缺文献中贡献了个人能力的调节效应, 积极调节PU和BI之间影响的研究。

研究创新

本研究通过纳入阻碍因素并探索个人能力在AR眼镜方面的调节作用, 为TAM提供了一种新的拓展。此外, 该研究还使创新公司能够通过用户的反馈来增强其产品和服务的设计。

Citation

Zaki, H.S. and Al-Romeedy, B.S. (2024), "Chatbot symbolic recovery and customer forgiveness: a moderated mediation model", Journal of Hospitality and Tourism Technology, Vol. 15 No. 4, pp. 610-628. https://doi.org/10.1108/JHTT-11-2023-0374