Will AI Help Call Centers?
Written: May 2023
Word count: 896, Reading time: 3.4 minutes
In 2001, I wrote Net Attitude: What It Is, How to Get It, and Why Your Company Can’t Survive Without It. The focus was to share with organizational leaders how they could use a “net attitude” to make their organization more successful by using the web to be more responsive to customers and constituents. The classic example of a poor net attitude is often found in one of the many call centers as I wrote more than 20 years ago. Many have still not improved. The pre-recorded voice greeting begins with, “Please pay attention because our menus have recently changed.” How is it the menus of all call centers have recently changed? If you press “0”, you sometimes get a response saying, “You have pressed an invalid key”. “Please enter your 16-digit account number” is often followed by a person in the call center right after you entered the 16-digit number asking, “What is your account number?” These annoying, often repetitive responses are not due to a technical problem. They are due to a lack of net attitude.
According to a 2022 survey by the American Customer Satisfaction Index (ACSI), online retailers had an overall satisfaction score of 79, with the top tier online retailers scoring in the range of 75 to 90. Telecommunications companies had a score of 68, with some much lower. One of the key elements in the difference is the call center.
I once had a question for Comcast technical support which illustrated their lack of a net attitude. I tried my best but eventually gave up reaching someone who could understand the problem I was having. It wasn’t the hold time; it was the attitude they deploy. I called the main number. The first prompt said to press 1 if the call was about USC. I have no idea what USC means. University of Southern California? I was next prompted and asked for my reason in calling. I said, “technical support”. I then listened to 15 seconds of clicking sounds which I was supposed to think was an agent typing my request on a keyboard. Do they think consumers are stupid and not realize the clicks were synthesized. Next, I was asked to press 1 for Slow Internet, 2 for Connection Problems, 3 for Wi-Fi Password, or 4 for email trouble. There was no other choice, and none of the choices offered related to what I was calling about. If I did not select one of those four choices, I could not proceed. I pressed “0” hoping to get to a person and the call center hung up. Cable companies face a number of issues related to pricing and contractual terms, but I believe annoying call center protocols contribute to the frustration and poor ratings.
According to Site Selection Group, a Dallas, TX, and Greenville, SC provider of global location advice, there are more than 7,400 call centers in the United States employing more than three million people. John McCormick at the Wall Street Journal, described how Cogito, a Boston based augmented intelligence company, is using voice AI to make call centers more effective. McCormick explained,
As calls come into a center, they are streamed to Cogito’s system, which evaluates hundreds of data points, including speech rate, tone and more. If agents are pausing before answering questions, it could indicate they’re distracted. If customers raise their voices, it could be a sign of frustration. When the Cogito system detects a possible issue with a call, it sends a notification in the form of an icon or short message to the staffer’s screen. It is a suggestion that the agent [can] recognize and acknowledge the caller’s feelings.
The Cogito system in effect coaches the call center agents to help them become more confident, engaged, and empathetic. McCormick quoted the Cogito CEO as saying, “Learning to speak to different customers is a real skill. You’re not born with it. You have to learn it.”
Some insurance companies have found the Cogito AI helpful in improving first call problem resolution as much as 10% and with increased customer satisfaction, but it clearly has a long way to go. The problem is not just empathy. The problem is call center people often do not have the knowledge to answer the questions. This will change dramatically over the next few years. We will find an AI will be able to learn how to satisfy customers better and faster than human agents. By applying machine learning to a large number of customer calls and matching the nature of the problems with solutions which solved previous problems will enable much higher accuracy than a human can achieve. We can look forward to AI-created voices which sound like a human and an AI which can understand the questions we ask without asking us to press 1 for this and 2 for that. I would not recommend working in a call center as a good career path. The three million call center workers are exactly the kind of jobs which AI will threaten.
Apple and others are using AI but it is not highly effective. The recent explosion in the use of generative AI will change that. The big tech companies are well aware their chat bots are not always accurate, and they are investing large amounts of money to improve them.