Emotional Intelligence in AI-Mediated Customer Service: Contextual Effects Across High- and Low-Spending Interactions

This study investigates how distinct dimensions of chatbot emotional intelligence (EI) influence customer satisfaction across high- and low-spending service contexts. Drawing on Goleman’s established emotional intelligence frameworks, the study conceptualizes chatbot EI through self-awareness, self-regulation, social awareness, and relationship management, and examines their differentiated and context-dependent effects using a controlled online experiment (N = 830). Results show that emotional intelligence is not a monolithic capability in AI-mediated service: self-regulation functions as a baseline hygiene factor, whereas social awareness, self-awareness, and relationship management significantly enhance customer satisfaction. Importantly, spending context operates as a boundary condition. In high-spending interactions, customer satisfaction is driven primarily by self-awareness and social awareness cues that signal emotional recognition and contextual attunement. In contrast, in low-spending interactions, relationship management cues emphasizing reassurance and interactional continuity are most effective. These findings contribute to research on emotionally intelligent AI by demonstrating the importance of contextual calibration rather than generic empathy. From a business-for-society perspective, the study highlights how responsible emotional design of AI can support more respectful, trustworthy, and socially attuned customer experiences without amplifying ethical risks associated with emotional manipulation.