Apologizing with a smile or crying face? Exploring the impact of emoji types on customer forgiveness within chatbots service recovery

While advancements in AI have facilitated the uptake of chatbots across a range of sectors, incidents of service failures have been documented in numerous instances involving chatbot users. In this context, it is of paramount importance for chatbots to adopt appropriate service recovery strategies in order to mitigate and minimise the negative impact of chatbots failures. This research proposes that the use of emojis by chatbots when apologising represents an effective strategy for the recovery of customers following the occurrence of online service failures. The results of three scenario-based experiments indicated that the use of negative emojis by chatbots was more likely to result in customer forgiveness than the use of positive emojis, provided that the severity of the service failure was low. Moreover, the utilisation of negative emojis by chatbots fosters customer forgiveness by enhancing perceived empathy, whereas the deployment of positive emojis has the opposite impact by increasing perceived ambiguity. These findings provide crucial guidance for online retailers in the design of chatbot customer service strategies, emphasizing the pivotal role of subtle emoji differences in attaining customer forgiveness.