By better understanding how emotion influences the willingness or ability of customers and employees to participate in service production, we can refine our ability to design operational processes that are both efficiently delivered and poised to improve engagement over the long term.

 

“The Negative Effects of Positive Feedback from Artificial Intelligence”

Companies often use standardized protocols to streamline service delivery, yet securing worker compliance remains a persistent operational challenge. This study investigates whether real-time feedback from artificial intelligence (AI-RTF) can improve adherence to scripts, reduce task completion times, and enhance job satisfaction—while also examining how its psychological effects on workers shape these outcomes. We contribute to research on human–AI collaboration by analyzing how AI-driven feedback influences productivity, anxiety, self-confidence, and engagement, relative to delayed feedback from human supervisors—underscoring the emotional dynamics of frontline service work. In a preliminary field experiment conducted during the COVID-19 pandemic with newly hired remote call center agents, AI-RTF increased script compliance and reduced average call duration by 15.6%. However, model-free evidence showed no net reduction in call length due to a post-hoc finding: agents receiving real-time feedback initiated more call interruptions, effectively offsetting time savings. In follow-up laboratory experiments, confirmatory feedback from an AI evaluator increased compliance but also heightened anxiety, lowered self-confidence, and introduced variability in task duration. Our results suggest that while AI-delivered real-time feedback can drive behavioral gains, it may also incur psychological costs that may ultimately erode throughput. Managers should carefully weigh the trade-offs between operational efficiency and worker well-being when deploying AI-based monitoring and feedback tools in standardized service environments.

Media Coverage

Insights at Questrom, December 7, 2020, “Who does AI help: Customers or Companies?”

 

“Mitigating the Negative Effects of Customer Anxiety by Facilitating Access to Human Contact” 

Prior research in social psychology has shown that when people feel anxious, they seek advice from others. Yet, companies that operate in high-anxiety settings (like financial services, health care, and education) are increasingly deploying self-service technologies (SSTs), through which anxious customers transact without access to human contact. These companies may therefore face a classic efficiency-service tradeoff where gains in operational efficiency through automation also hamper service outcomes by neglecting customer anxiety.  This paper, set in the high-anxiety domain of financial services, investigates the value of facilitating access to human contact during SST usage.

In a field experiment conducted within the context of a credit union’s SST loan-approval process, an invitation to connect with a human loan agent increases the customer uptake rate of approved loans by 24%. Subsequent controlled laboratory experiments, which explicitly investigate the linkages among anxiety, choice satisfaction, and firm trust during SST encounters show that compared to participants without anxiety, those who are anxious consistently report lower satisfaction with their choices and less trust in the firm, regardless of the source of their anxiety. When anxiety is related to the SST encounter, facilitating access to human contact dampens anxiety’s negative effects on choice satisfaction and, by extension, increases firm trust. We further observe that these benefits of facilitating access to human contact arise even though very few customers choose to access a person. Taken together, this research reveals an opportunity for firms to maintain operational efficiency in high-anxiety self-service settings without compromising customer experiences and service outcomes.

Awards

INFORMS Behavioral Operations Section 2019 Best Working Paper Award, First Place

POMS College of Behavior in Operations Management 2019 Junior Scholar Paper Competition, Second Place

 

Media Coverage (Subscriptions May be Required)

Business Insider, November 26, 2024, “Customer Service Hits a New Low”

Business Insider, September 4, 2024,  “The hot new status symbol: Phone Calls”

Washington Post, March 13, 2023, “Vanishing customer support is driving us all insane”

Vox, January 26, 2023, “The death of the customer service hotline”

AARP, September 10, 2021, “Chatbots May Deliver the Best Customer Care”

Harvard Business Review Webinar, September 25, 2019. Listen to the replay here

Harvard Business Review Magazine, July-August 2019, “IdeaWatch: How Necessary Is the Human Touch?”

Harvard Business Review (April 2019, Digital Article): ” Why Anxious Customers Prefer Human Customer Service”

HBS Working Knowledge Infographic (April 2019): “Can I please speak to an actual person?”