How and Why the Use of AI (vs. Human) Agents Influences Sales Conversion Rate

Artificial Intelligence (AI) and algorithms outperform humans in a variety of marketing tasks. In the B2C sales frontier, many people expect AI-based agents to be a fitting solution to routine sales tasks, such as purchase suggestions and upsells. But when an AI agent suggests a purchase, will it elicit the same purchase intention as a human agent’s suggestion? We examined this question at the consumer level in seven pre-registered experiments (N = 3,806) across diverse sales contexts, purchase categories, and both online and offline shopping environments. We found that consumers infer a clearer profit-driven motive from an AI agent’s purchase suggestion than they do from a salesperson’s, which decreases their purchase intention. Supporting this mechanism, the AI agent’s disadvantage in sales conversion was mitigated when purchase suggestions were broadcasted indiscriminately (instead of individualized), and when the suggestion was no longer linked to immediate firm profits. Alternative explanations including a general distaste for or low engagement with algorithms were ruled out. Our experiments document a robust negative main effect of AI (vs. human) sales agents on consumers’ purchase intention, and find congruent evidence supporting the proposed mechanism: cynical consumer inferences. For managers, our results suggest that tradeoffs will have to be made when deciding whether AI agents should be used in consumer-facing interactions. This mechanism also generates practically relevant predictions on when and why using AI sales agents can be relatively more beneficial