As companies chase efficiency, it’s no surprise that today’s customer service call is more likely to be answered by a bot than by the much-maligned, outsourced operators of the past.
Executives celebrate this shift with near-religious zeal. AI doesn’t ask for vacation days, sick leave, or overtime pay. It’s a one-time investment that trims expenses and inflates margins—pure gold for companies chasing acquisitions or juiced valuations.
No wonder “AI-enabled customer service” is now a prized bullet point in every 2025 investor deck. Watch any interview with U.S. commerce officials, chamber of commerce members, or senior executives, and you’ll hear the same refrain: AI in customer service is a triumph of efficiency and value.
But does this dream match reality?
From the lofty perspective of the C-suite, yes. But from the consumer’s side, the mirror opposite is closer to the truth. It’s striking how two groups of seemingly rational people can look at the same situation and walk away with polar-opposite conclusions.
To the average customer, the phrase “the death of customer service” comes to mind—usually delivered with more colorful language than can be repeated here.
By 2025, calling a help line has become almost universally dreaded. The old irritations—thick accents, scripted platitudes, and the infamous, “Is your computer plugged in?”—now feel almost quaint.
Today, customers are forced to spar with computers: patronizing bots with faux personalities, trained through endless machine-learning (ML) loops. These digital gatekeepers don’t just lack the human touch; they demand that you be more precise, more patient, and more accommodating. Before you even get that far, you’ll likely be redirected to a website 3 minutes into the call—after navigating more menu options than a high-end sushi restaurant. By then, even the most loyal customer’s patience is gone.
The CLEAR Experience
For years, CLEAR, the tech company that handles travel document verification at airports, was the poster child for a customer-first model. On the front end, its service felt effortless: friendly staff at airports, efficient check-ins, consistently positive interactions. Many loyal users even added family members to share the convenience.
But nothing is free—and that’s where the trouble began. One longtime customer described adding his wife to the plan multiple times, paying each time, only to have CLEAR’s system repeatedly fail. Despite storing her full profile—birthdate, email, even facial recognition—the system acted as if she did not have an active account.
It’s a reminder that even the best smiles at the counter can’t compensate for back-end failures. When the system itself collapses, no amount of warmth can save the experience.
AI-Powered Frustration
When our CLEAR customer called the help line, he was greeted not by a person but by CLEAR’s “AI-powered” customer service. Technologists cheer: Perfect, this will be faster than waiting on hold.
They’re wrong. Again.
Anyone who has tried in the last 2 years to get a simple answer to a simple question has been coerced into becoming an unpaid beta tester. We’re forced to feed data into ML systems that stumble through failure after failure until they eventually learn how to answer one narrow question. Then someone asks a slightly different one, and the cycle begins again.
No matter how “clearly” the problem was explained—duplicate charges, missing account linkages—the AI simply couldn’t grasp it. Instead, it spat out irrelevant answers, none of which solved the issue. Worse, there was no clear path to a human being. Oh, how we now long for the days of outsourced call centers. Flawed as they were, at least they picked up the phone.
Every customer knows this drill: You beg to speak to a representative; you say, “Operator”; you hammer “0” until your fingers ache. In this case, CLEAR’s AI offered two choices: Wait on hold or request a call back. When our customer chose to wait, the system rewarded him by simply hanging up each time.
Finally, in desperation, he selected the “call me back” option. The system promised a response in 5–7 business days. That’s not a service window—that’s a dismissal.
Even turning to email offered no escape. Replies took more than a week—usually followed by silence. For comparison, even airlines—in the middle of weather chaos with half the country’s flights canceled—manage to return calls in hours. CLEAR, despite its reputation for flawless airport operations, failed spectacularly where it mattered most: on the back end.