AI Was Supposed To Fix Customer Service: It’s 100x Worse | The Deep

This is the story of Dave, whose customer service experience was so bad, it almost killed him. Dave is 44. He works a desk job, and he has restless leg syndrome and chronic insomnia. It's not a life-threatening condition, but it can really unravel your life. After months of trial and error, Dave's neurologist finally finds medication that works. For the first time in years, David sleeps six hours at night. Then January hits. And his insurance resets. The pharmacy texts him, your prescription is delayed, action required.

He clicks the link. The virtual assistant greets him by name. Hi, David, I'm your care assistant. How can I help? He selects medication issue, then prior authorization, then yes, I still need this medication. The AI assistant replies, your provider must submit additional documentation. So, Dave calls his doctor's office. A recording answers. To get started, please use our online portal for fastest service. So, he logs in. The portal says, insurance denied the claim. Your provider must submit additional documentation. No kidding. That's what I called to tell you.

He calls his insurer. After menus, music, and a soothing voice telling him his time is valuable, he's routed back online. ...to another AI assistant. This one tells him, the prior authorization is pending. The pharmacy texts again, authorization expired. By now, David hasn't slept in four nights. His legs twitch constantly. He stares at the ceiling until 3 a.m. and drives to work half awake. He finally reaches a human representative. David asks who can fix it. She seems genuinely concerned, but answers, unfortunately, I don't have the ability to override the system.

He asks for her manager. She doesn't actually have one, but she offers to redirect him to the main menu. No one denies him care. The one real person he talks to cares. The provider blames insurance. Insurance blames pharmacy policy. The pharmacy blames prior authorization software. The software says the workflow is complete. This story does have a happy ending. David drives to his local pharmacy, stands at the counter, offers cash, and happens to get a pharmacist willing to make it happen, until the next time his insurance resets.

David's story is far beyond frustrations we might have with returning a large screen TV or a messed up payment plan on Klarna. At this point, the AI machine's stunning stupidity is actually ruining people's lives. In what looks on paper like an age of abundance, with pharmacies stocked to the brim and doctors trained to the T in life-saving medications and therapies that the emperors of Rome could only dream of, We can't get those resources to the people who need them. And if we do, it's only after days and hours of wasted time and needless suffering.

No wonder the rage against the machine is real. In David's case, what looked like procedural efficiency became cruelty in real life. Not because anyone wanted to deny him care, but because no one retained the agency or had the authority to step in and fix the problem. The system was working as designed, which is precisely the problem.

It's demeaning, not just for Dave, but also for his doctor, his pharmacist, and even the nice Filipina lady at the call center. Customer service is teaching us to despise our own uniqueness, turning us, as Ian McGilchrist puts it, into identical, interchangeable pawns in a game they do not understand they are playing. An excellent project whereby the rich get richer and the poor get poorer. Call centers and AI efficiency are supposed to give us time back, and now there are fewer people involved in every previously human interaction, and yet we are all more exhausted and angry and stressed.

In his essay, Can You Still Be Human?, McGilchrist says, leisure is not ours, because while AI saves the super rich even more money by allowing them to do away with people in their businesses, this doing away with people in their businesses is not ours. People imposes a huge burden of wasted time on the rest of us, and your time is valued at precisely nothing. Who does not now find themselves constantly busy? The young, the middle-aged, and the old feel it alike.

We are all time poor, and time is not worth nothing. Time is life. Everything is now freighted with bureaucracy, and that bureaucracy is in what McGilchrist calls an unholy alliance with AI. Together, they colossally waste our time and drain us of life. So, I went down that rabbit hole. How did we get here, and how do we get out?

Big companies are moving fast, recklessly fast, to embrace artificial intelligence. And the tone is not caution or humility. It's giddiness. The same energy Silicon Valley always brings when it realizes it might finally eliminate troublesome costs, usually in the form of human beings. The unspoken message is clear. We won't need human beings anymore. For large corporations, this all sounds like salvation.

AI is sold as a triple win, faster service, lower labor costs, and, this is key, no decline in customer satisfaction. The pitch is always framed as neutral efficiency. But that neutrality collapses the minute you look at where the savings actually come from. Google's executives are positively euphoric about it. According to internal briefings and reporting on Gemini's latest rollout, executives openly describe a future where AI agents handle entire workflows, sales, support, logistics, without human bottlenecks. Entire departments become optional. This has been going on for a while.

Best Buy is a perfect case study in what this looks like in practice. Not long ago, every Best Buy location employed MCSAs, multi-channel sales associates. These were trained employees whose only job was answering phones. They knew the local inventory. They could walk on the sales floor, physically verify stock, and give a real answer to the customer. They were accountable because they belonged to your store. Around 2017, that system was scrapped.

Phone support was centralized into national call centers. Fewer people, cheaper labor, higher efficiency. On paper, the savings looked obvious, but it was actually a race to the bottom. Best Buy is just one example of the shift. For most of the 20th century, customer service was local. Human and relational. You spoke to the receptionist at your doctor's office. You knew the bank teller. Store managers lived in your community and had authority to solve problems. Service was personal and accountability a feature of the system.

Then, scale changed everything. As corporations went national and global, customer service became a cost center to be minimized. Call centers replaced front desks. Scripts replaced judgment. Offshoring wasn't just a manufacturing phenomenon. It was also quickly adopted as a service practice. Outsourcing to massive hubs in Asia and Latin America promised efficiency and lower labor costs, but at the price of distance. Cultural disconnect and worker burnout. Customers became ticket numbers. Workers became interchangeable voices measured by handle time. When we went global, something crucial disappeared.

Local knowledge and human ownership. Today, the person answering the phone can't see the shelves, doesn't know the products, and has no incentive to care whether you ever come back. That's the hidden pattern of globalist AI enthusiasm. No matter how many times they say it, growth is not about making things better for the consumer. It's about creating a scalable bureaucracy to make more consumers and to make more money. The race to the bottom is a feature, not a bug, of scalability. At a certain point, the bureaucracy becomes so big that it liberates itself from responsibility to individual consumers.

And consumers? Well, we participate, too. Of course we do. I like returning Amazon packages from my couch. I like tapping a button on Christmas Day and watching a refund appear without waiting around in line at a store. listening to elevator music on hold. Convenience is seductive, especially when we're already exhausted and time poor. None of this requires corporate executives to hate customers. They don't need malice. They're simply optimizing to win, to stay rich, get richer, and reassure themselves that cutting humans out of the equation is still somehow for the customer.

But listen closely. That word customer keeps appearing. Satisfaction, speed, frictionless experience. And yet, when you ask actual customers how it feels, they hate it. Actual human customers despise AI-driven customer service. Not mildly, not theoretically, measurably. Survey after survey tells the same story. Nearly 70% of customers say, They would leave a company, if they could, after one bad AI customer service interaction, not three bad experiences, one. Why? Because AI doesn't just fail. It fails in ways that feel dehumanizing. Author and customer experience expert Amas Tanuma explains it this way.

Americans are incredibly gracious when they start. If it's on a scale of 1 to 10, most people start at a 9 or a 9.5, but then you start this interaction, and you're met with an automated system. Press 1, press 2, or a machine you're trying to communicate with. Automated menus, voice recognition that can't understand accents. or frustration, chatbots that request information you've already provided, and finally, if you're lucky, a human being who asks you to repeat everything from the beginning. By then, that nine has collapsed into a four, and God help you if they transfer that call.

Now your graciousness levels are at a one, if they survived the ordeal at all. The Verizon CX Annual Insights Report Surveying 5,000 consumers across seven countries found an astonishing 28-point satisfaction gap, 88% satisfaction with human agents, just 60% with AI interactions. Harvard Business Review reports that 77% of people find chatbots frustrating, and 88% still prefer human agents, even when the AI systems perform with near-perfect technical accuracy. That last point matters. This isn't about bugs. It's about psychology. Humans despise being trapped in systems that can't recognize exceptions.

AI doesn't understand context. It doesn't bear responsibility. And worst of all, it often pretends to be human while offering no human accountability. Former Google chief decision scientist, Cassie Kozarkoff, put it plainly, AI makes mistakes. AI can't take responsibility for those mistakes, and users hate being tricked by a machine posing as a human. In other words, people would rather wait longer or even receive an imperfect answer from a real person than get a fast, wrong response from a bot wearing a human mask.

And yet, companies continue to double down. Why? Because the calculus changes when customers Can't leave. Enter the monopoly. Try opening a bank account without the internet. Impossible. Try finding a major healthcare provider that doesn't route you through AI triage systems, digital portals, automated billing workflows, and bot-mediated authorizations. You can't. You can delete Klarna, and you might escape Amazon, but you cannot escape money or medicine. That's where customer frustration morphs into something darker. The 2025 National Customer Rage Study shows the emotional cost.

77% of customers experienced a serious product or service problem last year, more than double the rate in 1976. 64% of them felt rage. 50% complained, either online or by trying to call a number. 32% reported emotional stress as a direct result of the resolution process. Digital complaints now dominate, and yet 43% of them disappear into algorithmic voids, never acknowledged. Revenge seeking behavior among customers has tripled since 2020, and I kind of understand why. This isn't impatience, it's despair. We were promised speed and leisure, and instead, We got bureaucracy without humans and without any regard for our time, money, or sanity.

And when AI governs systems that people cannot opt out of, their healthcare, insurance, finance, that despair becomes dangerous. That's when customer rage feels like a prisoner's dilemma. Well, I don't know about you, but at this point, my blood is on boil. I am a little worked up. But what do we do? Remember Ian McGilchrist from the start of this video? He talked about how time is life, and he also gives us one key to solving our prisoner's dilemma. Resist fatalism. The quiet shrug, well, this is just how it is now, is not realism.

It's surrender. And at scale, it's suicidal. The seductive promise of the entire industrial revolution, and now the AI revolution, was always the same. We would save time, grow richer, and gain leisure. None of this happened, unless you count the fact that large screen TVs are ridiculously affordable now. This is the deeper betrayal. We all feel it, and we must resist. Something that 10 years ago could have been resolved by a five-minute phone call with another human now requires hours of going around in M.C.

Escher-like loops online, platform to platform, portal to portal with no exit. Nothing on offer fits real life. Will those growing up today even be able to recognize how human business and markets could be, how humane they once were, if there were more people involved?

We have to use our moral imagination to conjure a vision and concrete plan for pursuing more humane ways to conduct our lives and business. Can you shop local? Pick a family-owned restaurant over a chain. Choose local entertainment like absurdly bad community theater over absurdly bad Marvel remakes. Buy your kids one Christmas gift from the mom-and-pop Etsy store over six gifts from Walmart. It might feel like all your little efforts to buy your neighbor's eggs or barter for music lessons, choose the car shop with the family name on it over Meineke, are in vain.

But you are actually carving out a zone of freedom for yourself and your family that is sacred. Spend as little time on your devices as your life permits, so that when the machine does come for you, like it came for Dave. You can face it with more calm and peace. Companies should also take note. Trust is the real currency here, and right now, it is the most scarce resource in the economy. Sure, AI will save monopolies millions of dollars, but they will do so by sacrificing trust, dignity, and human presence, gambling that customers will always have nowhere else to go.

History tells us that's a bad bet. Systems don't last forever. Empires don't either. And when things start to crack, when people finally reach their limit, we the customers won't be looking for faster bots or sleeker platforms. We will be looking for humans. And we already are.

If you liked this episode, be sure to subscribe to LOOPcast. We bring you a new episode of The Deep every Thursday at 4 p.m. Eastern.