Determining the causes of user frustration in the case of conversational chatbots

Artificial intelligence (AI) is expected to transform many facets of our existence with the gradual penetration of the semantic web. One such AI-based technology is a chatbot that considerably alters numerous business verticals to enhance users and businesses. Even though companies are extensively contemplating using chatbots to supplement or even replace humans, users have exhibited anger, confusion, and customer dissatisfaction with chatbots. The current study aims to find the reasons for user frustration in the case of conversational chatbots and adopts exploratory research using a critical incident technique (CIT) survey. Results are analysed using sentiment analysis and topic modelling with Python. Results of the study indicate four significant causes of user frustration, i.e. poor semantic understanding, lack of customisation, lack of Humanisation of the bot, and poor competency. These causes have further been examined across various socio-demographic variables and industry contexts. This work would theoretically advance the bot literature and fulfil the gap by understanding the reasons for frustration. This would further help bot developers and marketers in designing and training chatbots that lead to better engagement and enhanced customer experience.