聊天機器人服務失誤對於顧客之影響

Abstract

近年來，隨著聊天機器人在顧客服務中的應用快速增長，商家導入聊天機器人所帶來的顧客體驗與潛在影響逐漸受到重視。雖然已有研究關注聊天機器人的互動設計、接受度與顧客滿意等議題，但針對服務失誤情境下，顧客是否會產生與聊天機器人相關的焦慮情緒，進而影響其對服務品質的評價與後續行為意圖，仍缺乏系統性的探討。為填補此重要研究缺口，本研究旨在建立一實證模型，從商家聊天機器人的服務失誤出發，探討其是否會導致顧客針對聊天機器人的焦慮情緒，降低顧客感知服務品質，進而影響顧客後續行為意圖。

本研究以具有聊天機器人使用經驗、並在六個月內有經歷過商家或品牌聊天機器人服務失誤，且對整體事件及使用過程體驗都還保有一定程度之印象的使用者為調查對象，透過SurveyCake網站進行電子問卷調查，採用簡單隨機抽樣法回收有效問卷共511份，採結構方程式（Structural Equation Modeling）的統計方式進行模型分析與假說驗證，研究結果發現，經歷過商家聊天機器人服務失誤（包含功能性失誤、互動性失誤），會正向影響顧客的聊天機器人焦慮，進而降低顧客感知服務品質，並對顧客的後續行為意圖造成負面影響。

綜合上述，本研究表明，商家的聊天機器人功能性失誤與互動性失誤，是顧客聊天機器人焦慮的重要影響因素。本研究亦針對次研究結果進行學術與管理意涵方面的探討，並提供未來研究建議與發展方向。

In recent years, with the rapid growth of chatbot applications in customer service, the customer experience and potential impact of chatbots introduced by businesses have gradually gained attention. While existing research has focused on issues such as chatbot interaction design, acceptance, and customer satisfaction, there remains a lack of systematic exploration into whether customers experience anxiety related to chatbots in service failure scenarios, which may influence their evaluations of service quality and subsequent behavioral intentions. To address this critical research gap, this study aims to establish an empirical model that examines whether service failures in business chatbots lead to anxiety toward chatbots, reduce perceived service quality, and ultimately influence subsequent behavioral intentions. This study targeted users who have prior experience with chatbots, have encountered service failures from business or brand chatbots within the past six months, and retain a clear impression of the overall event and usage process. Data were collected through an online questionnaire distributed via SurveyCake, using simple random sampling, resulting in 511 valid responses. Structural Equation Modeling (SEM) was employed for model analysis and hypothesis testing. The results indicate that experiencing service failures from business chatbots—including functional failures and interaction failures—positively influences customer chatbot anxiety, which in turn lowers perceived service quality and negatively affects subsequent behavioral intentions. In summary, this study demonstrates that functional and interaction failures of business chatbots are key antecedents of customer chatbot anxiety. Theoretical and managerial implications are discussed, along with suggestions for future research and development.