When bots’ empathic expressions backfire: Exploring the negative moderation between chatbot empathic expressions and monetary compensation for angry customers

Abstract This study examines the impact of empathetic words from chatbots on perceived service quality and attributions of chatbot empathy, particularly when combined with monetary compensation (e.g., discount coupons) for customers upset by service failures. We hypothesize that empathetic words from a chatbot positively affect both service quality and attributions of empathy, all else being equal. However, we also propose that empathetic words negatively moderate the effect of monetary compensation. Specifically, while compensation alone improves outcomes, its effect reverses when paired with empathy. To validate this, we designed four scenarios in which chatbots varied in the degree of empathetic words and coupon offerings when addressing dissatisfied customers. Data from 610 participants confirmed that although empathetic language enhances perceived service quality, it transforms the positive effect of monetary compensation into a negative one. This study offers two key contributions. First, it reveals the negative interaction between empathy and compensation. Second, it provides a nuanced conceptualization of chatbot empathy through precise experimental manipulations, offering new insights into how empathy functions in automated customer service interactions.