Artificial persuasion: profit motives and the persuasiveness of AI sales agents

As companies increasingly delegate sales tasks to AI-powered agents, an important question emerges: How effective are AI agents at persuading potential customers to make a purchase? Across five experiments spanning three selling contexts, we find that customers’ responses to AI sales agents depend on how clearly the situation signals that the agent is motivated by the company’s financial interests. When profit-motive signals are strong (e.g. the agent recommends a more expensive product), customers are willing to pay less for products recommended by AI than by human agents. When those signals are weak, the gap closes or reverses. We show that perceived benevolence—the belief that the agent wants to act in the customer’s interest rather than the company’s—helps explain this pattern, because customers doubt that AI can override its training to prioritize them over the seller. Our work contributes to research on persuasion knowledge by identifying a novel asymmetry in how customers apply it to AI versus human agents. This work also contributes to research on trust in AI by drawing attention to benevolence, an under-studied dimension of trust that influences customer behavior when ulterior motives are present.