I am on hold with music that is no doubt the same music played on loop in the fifth circle of hell. It’s been 35 minutes. It will be 50 minutes before the system will hang up on me because I am not a leaf in its automated customer-service decision tree. I am angry and frustrated. There is no way to talk to an actual human.
My credit card company referred me to a monitoring service when its data systems were breached (again). The monitoring service sends me a sexual-predator alert about my file. What? I can’t get into my account because my password won’t work, and I have to be on the site to change my password.
I call the customer-service number and try to get myself routed to a human by saying the word “agent” 25 times. By the time I get to an actual human, she sympathizes and says that the magic word is “fraud”—you always get to a human immediately because in the case of fraud, the credit card company is the victim. The fraud humans can route you to other humans who can help me change my password. It turns out that no one thinks I’m a sexual predator, I just live in the vicinity of one.
Sound familiar? In the increasingly automated world of customer service, custom service—actual humans who can solve problems and answer questions—is on the wane. What we increasingly get are automated customer-service systems that are frustrating to deal with and don’t solve our problems. Most people call customer service only when something is not working, but it seems that these systems are designed to address “common” problems—problems that you don’t necessarily have. When there are no people to be found to explain things to, you may end up in an endless loop, having a bot tell you the same annoying irrelevant thing over and over. How is that customer service?
In my own fantasy world, customer service is custom service—service that targets my needs and my problems, suggests good solutions, and assists me in implementing them, all via a human who can take into consideration the tangible and intangible contexts relevant to a remediation that fits my needs.
Sounds dreamy, doesn’t it? Such solutions actually exist in the real world. For instance, in Northampton there is a store that sells bras. When you go there, there are expert humans who work with you to find and fit bras that work for your body. I understand that this is an experience that half of the people around the world (and many of my readers) will never have, but trust me, it’s wonderful. The experts at the store help triage a myriad of choices of bras: different colors, cup sizes, wired and unwired styles, different fabrics, and other options. They listen carefully as you tell them about the kinds of clothes you wear, your lifestyle, and your preferences. In the end, you can get something that really works for you and enjoy the experience of getting there. That experience is way different than choosing something on Amazon and hoping it will fit.
The bra store is all about custom service rather than homogenized, automated customer service. Custom service works there because the bra store is not a mega-company with millions of customers. Plus, we’re probably a ways away anyway from automated systems that will find and fit your bra.
My experience with automated customer service is that it homogenizes individuals and the problems they are likely to have and separates them from the people who can actually help.
Look, I understand that it’s harder and more expensive to do customer service at scale—when you have millions of customers—and that there’s no cost-effective way to hire enough people to give everyone a personalized experience. But there’s something else going on, something we should not lose in cyberspace: a real sense of respect for your customer and an understanding that without satisfied customers, businesses can cease to exist.
You can do custom service at scale with millions of customers. Some companies do this well and they invariably earn my loyalty. When I have a question about my Fidelity account that I can’t answer by going to the website or chatting with the chatbot, I can call Fidelity. Within a relatively brief amount of time, the bot who answers the phone verifies my voice and connects me with a human who can track down the answer to my question. Something similar happens when I call USAA Insurance. In 2025, USAA was in the Fortune 100 and Fidelity was the third-largest mutual fund company in the U.S.
Apple also has a phone number that connects me with real humans who can help solve problems. In addition, they have brick-and-mortar stores with a “Genius Bar” staffed by actual humans who apparently are not allowed to make me feel stupid. (OK, most of these folks are not geniuses, but they understand Apple products and systems way better than I do.) You probably have your own list of companies that make it easy for you to deal with them and easy for you to solve problems. My guess is that all or most of them have knowledgeable humans available early and often.
Digital technologies should help us do things better, not worse. Algorithmic efficiency is not a substitute for human empathy and judgment. Can’t we create a cyberspace that has both?
Reprinted from Better Tech: Putting People First in Cyberspace by Francine Berman with permission from MIT Press. Copyright 2026.
The post Why AI Makes Customer Service Worse, And What Companies Can Do To Fix It appeared first on TIME.
