Research Highlights

Short-text social media platforms such as Twitter (now X) and Weibo are popular channels for customer–brand interactions. When posting their queries or complaints, customers often express emotions such as frustration, anxiety, and gratitude. Brands recognize the need to address them with empathy. Traditionally, this function is fulfilled by trained support agents who combine practical solutions with emotionally attuned communication.

Recent advances in machine learning have enabled automated systems to generate empathetic responses by detecting customer emotions and conditioning replies accordingly. While these approaches show promise, they mainly focus on identifying emotion categories rather than the strength or explicitness of emotional expression.

Based on social psychology, this study posits that effective empathy in customer-brand interactions necessitates aligning the emotional intensity and explicitness of responses with customers' expressed emotions. By integrating emotional explicitness into machine learning models, this research aims to enhance the quality of AI-generated customer support across various industries while offering practical guidelines for the development of empathic chatbots.