Managing The Limits of AI For A Remarkable Customer Experience

My first trip to New York’s Saks on Fifth Avenue featured a super friendly elevator operator sitting on a small stool pushing a button to drive the elevator to the next floor and announcing the department as the elevator door opened. I exited at “Men’s Wear.” Once there, it was my first-time meeting Ella, a passionate salesperson specializing in men’s suits and related apparel. I recalled wondering how long before the nice elevator operator would be replaced with self-service buttons. The elevator operator was gone on my next trip to the Big Apple to see Ella.

The role of artificial intelligence in customer experience begins with the relevance of the words themselves. As customers, we clearly want our service to be an intelligent interaction, but do we also want a valued, authentic relationship? Or does it depend on the context? We don’t need nice elevator operators, no matter how smart they might be on the contents of each department store floor. But our “Ella’s” are a different story.

The race to out-AI each other is uncovering the need for a deeper understanding of customer expectations. We tout the features of ChatGPT as a necessary tool because it is accurate, fast, and important. We sometimes ascribed to it capacities that can seem magical and mysterious. We are awed that it can pass a state bar exam and write a perfect term paper. But it might also advise putting tomato, okra, or eggplant in a fruit salad (yuk!) because it knows they are all three fruits, not vegetables. What is the chief role of AI, and what are its limitations? Here are four features of a remarkable customer experience and the capacity and limits of AI for delivering each.

Know Me to Demonstrate Understanding

Customer understanding takes far more than a mastery of information and intelligence, it also requires insight. While AI can provide rems of timely customer data, it lacks the ingenuity and curiosity to demonstrate true understanding. Our head can be fooled by an AI-generated essay, but our heart cannot be fooled by AI-simulated empathy.

When the manager of a hotel was dissatisfied with the learnings gained from his customer intelligence sources, he held “free breakfast” focus groups with the taxi and Uber drivers who frequented the hotel property. He gained valuable insights by hearing what passengers said in the backseat. For example, towels that smelled slightly scorched did not just mean they remained too long in the housekeeper’s dryer; it signaled fear of a hotel fire started in housekeeping. Dust accumulation under the bed meant bugs, not just an unkept hotel room.

MORE FOR YOU
TikTok Ban: App Will ‘Go Dark’ Unless Biden Administration Intervenes, Company Says—Here’s Everything We Know
New Hacking Disaster Warning For Gmail, Outlook, Apple Mail Users
FBI Warns iPhone, Android, Windows Users—Do Not Install These Apps

Enlighten Me to Deliver Wisdom

Customer wisdom means more than the superficial knowledge needed to follow instructions; it means the level of deep learning that enables intelligent decision-making and astute planning. AI can be a great purveyor of teaching. However, it takes a mentoring relationship to ensure confident competence. Learning is a door opened only from the inside. That welcoming of learning requires rapport borne of authenticity and compassion; features AI does not possess.

My wife’s hairdresser occasionally gets a permanent. His reason? “Permanents can be embarrassing and awkward for many women,” he told me. “I believed if I could go through exactly what my customers did, I would learn features of their experiences which I could improve.” AI cannot learn through empathy nor instruct with humility.

Unburden Me to Promote Peace of Mind

Most customer experiences are at risk of having hiccups and barriers that rob customers of the sense of security they value. Research shows that trust is the number one feature of customer expectations. Some gremlins can be unraveled with linear problem-solving, a prowess possessed by AI. However, some customer experience problems are more complex requiring inventiveness.

When a new office building opened, tenants complained about the slow or insufficient elevators, triggering a long wait at times of peak usage. The AI solution was to add an elevator, speed up existing elevators, or request tenants stagger the times they arrived or departed the office building. All these solutions were impractical for the owners of the office building. A solution required resourcefulness. The company installed mirrors in the lobby, causing tenants to become so preoccupied with watching themselves and others in the lobby mirrors they stopped noticing the wait. AI can do logic and predictability; only humans can do illogic and serendipity.

Enchant Me to Ensure Certainty

Customer confidence in an organization comes partly from a solid belief that the organization will be sufficiently adaptable to recreate itself as needed to remain competitive. It is signaled by pleasant customer surprises—evidence of experimentation and risk-taking. Value-added has been the traditional way to surprise customers—providing customers with more than they expected. However, there is a limit to constant addition since it can quickly become a standard part of a customer’s ever-increasing expectations. Value-unique, on the other hand, relies on “different than,” not “more than.” The bottom line is this: there is a limit to generosity but no limit to ingenuity.

Value-added, for example, is providing a complimentary bottle of cold water when a customer brings a vehicle to a dealership for routine maintenance. Value-unique might be ensuring a couple of Hazelnut K-cups (the customer’s favorite flavor) at the Keurig machine in the waiting area. AI can identify an array of value-added tactics, but it lacks the capacity for the whimsy and fun needed to craft value-unique surprises.

What Do We Do With AI?

“While AI is an efficient and revolutionary technology,” reported a recent IBM research paper, “It can lack the human touch that customers might be seeking. A customer likely still expects some level of human engagement and empathy. Overusing AI might result in customers feeling disconnected.”

Successful organizations recognize that AI can be a valued partner if provided the support needed to do well what it can do well—provide helpful customer information, conduct customer teaching, provide linear problem-solving, and suggest ways to make the customer experiences more value-added. Partnership is a win-win relationship grounded in continual learning, acceptance, and encouragement. The more we learn about AI and its gifts, the better we can set it up for mutual success. Collaboration, not fear-laden caution, can fuel a connection that opens an opportunity for achievement and progress.