Three Things to Know About Customer Resistance to AI

Three critical insights emerge about customers' reluctance to engage with AI customer service tools. First, customers avoid chatbots primarily due to two interrelated factors. In a simulated customer service scenario, participants faced a choice between two options: one requiring a wait time for resolution, and another offering instant service but with occasional errors routing them back to the wait queue.

Despite the time-saving advantage of the instant service, participants chose the wait option 28% of the time. This phenomenon, dubbed "gatekeeper aversion," stems from both the uncertainty and multi-step nature of the process, regardless of whether it's handled by a human or AI. When presented as a chatbot instead of a human, this reluctance increased by an additional 10 to 20 percentage points, a phenomenon labeled "algorithm aversion."

Follow-up experiments suggest that providing transparency about a chatbot's limitations and displaying the expected wait time for each option could improve adoption rates.

Second, the medium through which bad news is delivered significantly impacts customer acceptance. Research shows that customers are more likely to accept unfavorable offers from AI than from human agents. In one study, 78.6% of customers accepted a subpar resale price from an AI, compared to 60.4% who accepted the same offer from a human.

Conversely, when the offer was favorable, human agents saw a substantially higher acceptance rate of 89%, versus 76% for the AI. This discrepancy arises because individuals do not attribute human intentions to AI, making it less susceptible to being viewed as "selfish" when providing subpar service, nor as "generous" when offering unexpected improvements.

The effect is particularly pronounced when the AI is portrayed as highly mechanistic; a more human-like persona diminishes its advantage in delivering bad news.

Lastly, a straightforward two-question assessment can accurately predict whether customers will embrace or reject AI tools. A meta-analysis of 163 studies involving over 82,000 participants identified two key factors influencing customer preference for AI: perceived capability and the need for personalization. When AI is perceived as more capable than humans at a given task, and when that task does not require personalized attention—such as in sales forecasting or playing chess—customers lean towards AI solutions.

However, in scenarios where personalization is essential, such as when forecasting individual customer sales or providing tailored customer support, customers overwhelmingly prefer human interaction. Before implementing AI to handle customer-facing roles, leaders should carefully evaluate the AI's capability relative to customers' expectations for personalized service.