Maskin med Medkänsla? : En kvantitativ studie om chatbotars emotionella intelligens vid hantering av servicefel

Background: The use of chatbots has increased dramatically in customer service, but many consumers remain skeptical about the technology. A key challenge is how chatbots manage service failures, where emotional intelligence may play a critical role. Previous studies have primarily focused on human agents, leaving a research gap regarding how emotional intelligence functions in chatbots and how this affects customer reactions, especially at different severities of service failure. Purpose: The aim of this study is to examine the role of a chatbot’s emotional intelligence in reducing customers’ tendency to spread negative word-of-mouth. The study also intends to investigate how perceived warmth and competence mediate this relationship, as well as how the severity of failure moderates these effects. Method: The study is based on quantitative methodology with a deductive approach. An experimental design is used where 566 respondents are randomly distributed across four scenarios combining high or low emotional intelligence and severity of failure. Furthermore, data is collected through surveys and analyzed using statistical methods such as t-test, Hayes PROCESS model and ANOVA. Conclusion: The study shows that chatbots with high emotional intelligence reduce customers' tendency to spread negative word-of-mouth after a service failure. The effect is due to the fact that emotional intelligence increases the perception of the chatbot's warmth, while competence does not have the same impact. Furthermore, these results remain consistent regardless of the severity of failure. Therefore, emotional intelligence should be seen as a key attribute in the development of chatbot systems.