Predicting Consumer Responses to a Chatbot on Facebook.

As chatbots have become increasingly popular over the past years, most social networking sites have recognized their far-reaching potential for commercial purposes. Their rapid and widespread usage warrants a better understanding. This study examines the effectiveness of chatbots on Facebook for brands. The study proposes and tests a model based on the Consumer Acceptance of Technology model (CAT-model) including three cognitive (i.e., perceived usefulness, perceived ease-of-use, and perceived helpfulness) and three affective (pleasure, arousal, and dominance; PAD-dimensions) determinants that potentially influence consumers' attitude toward brands providing a chatbot, and hence, their likelihood to use and recommend the chatbot (i.e., patronage intention). Structural equation modeling analyses show that two cognitive (i.e., perceived usefulness and perceived helpfulness) and all three affective predictors are positively related to consumers' attitude toward the chatbot brand. The findings further indicate that attitude toward the brand explained a significant amount of variation in consumers' patronage intention. Finally, all the significant determinants also have an indirect effect on patronage intention, mediated through attitude toward the brand. In conclusion, our findings hold valuable practical implications, as well as relevant suggestions for future research.