Emotion in consumer–chatbot interactions: a systematic literature review from a service marketing perspective

Abstract

Chatbots increasingly handle frontline service encounters in which consumers seek help, resolve problems, and evaluate providers, making consumer emotion important for understanding automated service experience. This review examines which service-encounter claims empirical consumer-chatbot emotion research can support and how future work can connect those claims to service marketing theory and practice. Following SPAR-4-SLR and PRISMA 2020 reporting, the review synthesizes 74 empirical articles through the theories-contexts-characteristics-methodology (TCCM) framework. The literature has grown rapidly but remains concentrated in routine transactional episodes, with service failure, recovery, and relationship-maintenance episodes receiving less coverage. Measurement practices commonly rely on elicited post-encounter responses, limiting access to emotional dynamics within service episodes. Construct use varies across chatbot-expressed cues, consumer attributions about the agent or firm, and consumer-felt emotion, making construct location central to interpretation. Encounter type provides the review’s organizing contribution by connecting these patterns of construct use and measurement to episode goals, emotional stakes, and service outcomes across routine, recovery, and relationship-maintenance episodes. An encounter-contingent scaffold and propositions link episode conditions, chatbot cues, consumer appraisal, consumer-felt emotion, and service outcomes, thereby structuring a future research agenda. For chatbot service design, emotional signaling becomes an episode-specific service-design choice rather than a generic marker of human-like interaction.