Customer service chatbots: Anthropomorphism and adoption

Firms are deploying chatbots to automate customer service. However, miscommunication is a frequent occurrence in human-chatbot interaction. This study investigates the relationship between miscommunication and adoption for customer service chatbots. Anthropomorphism is tested as an account for the relationship. Two experiments compare the perceived humanness and adoption scores for (a) an error-free chatbot, (b) a chatbot seeking clarification regarding a consumer input and (c) a chatbot which fails to discern context. The results suggest that unresolved errors are sufficient to reduce anthropomorphism and adoption intent. However, there is no perceptual difference between an error-free chatbot and one which seeks clarification. The ability to resolve miscommunication (clarification) appears as effective as avoiding it (error-free). Furthermore, the higher a consumer's need for human interaction, the stronger the anthropomorphism - adoption relationship. Thus, anthropomorphic chatbots may satisfy the social desires of consumers high in need for human interaction.