Hmm, the effect of AI conversational fillers on consumer purchase intentions

AI conversational agents are increasingly prevalent in marketing practices. A recent effort to make AI conversational agents humanlike is the use of conversational fillers, such as uh, um, and hmm. While computer science research has shown that conversational fillers used by AI agents generally improve users’ interactional experiences, their effects in marketing remain unexplored. This research examines AI conversational fillers in a sales and promotion context and shows that they trigger consumers’ suspicion of ulterior motives, thereby decreasing their purchase intentions, an effect moderated by organization type (for-profit vs. nonprofit). We conducted one field experiment and four online experiments with different modalities, products, and languages to provide empirical support for the hypotheses. In so doing, this research contributes to the research on chatbot natural language and AI persuasion by showing that the persuasion knowledge model can come into work in consumers’ interaction with AI agents.