Imagine a company announcing its new AI-powered customer experience initiative. The exchange with its customers goes something like this:
Company: “Our intentions are honorable. We desire above all things to help. We will not force anything on you. You will take only that which you choose to take.”
Customers: “Precisely what are your motives? Are we to presume there is no ulterior motive?”
Company: “There is nothing ulterior in our motives. Nothing at all. You will discover this for yourselves before too long, simply by testing the various devices which we will make available to you. We ask only that you trust us. Only that you simply trust us.”
Does this sound familiar?
You may actually have heard these exact lines before. But they’re not from a company talking about AI. They’re from a 1962 Twilight Zone episode in which aliens arrive on Earth bearing gifts for humanity and submit to United Nations questioning about their motives.
To prove their case, the aliens even leave behind a book with a title that US government code breakers translate as “To Serve Man.”
The humans are satisfied with the answers they get — which ridiculously come from a lie detector machine for humans — and they build peace and prosperity with the aliens’ devices. Then the aliens suggest reciprocal visits between their two planets, and the humans board their ships en masse. What could possibly go wrong?
***
“Too many businesses are deploying AI to cut costs, not solve problems, and their customers can tell the difference.”
That’s a real quote about AI from the Qualtrics 2026 Consumer Experience Trends Report based on a survey with United Nations of customers (20,000+ people across 14 countries).
Thank goodness the people of Earth are more aware of what’s happening than our Twilight Zone counterparts were.
That’s a good thing (a real good thing) for us humans.
It should stop every executive cold.
What the boardrooms leave out
The customer service problem with AI isn’t about the technology itself. (Customer pain points hardly ever are.) The problem is that the technology is working exactly as designed — which was most likely not with customers at the center.
72% of CEOs have taken personal ownership of AI decision-making — double the share from a year ago — and half believe their jobs depend on getting it right. The primary framing for CEOs is ROI, productivity, and cost reduction. Speed to deployment is their metric.
Somewhere between the board mandate and the deployment timeline, the CEOs skipped a step. Customers noticed.
81% of people believe companies are deploying AI to save themselves money, not to improve service. Yup, we’re a perceptive bunch. We’re experiencing the difference between AI designed around organizational convenience and AI designed around human value — and we’re well aware which episode we’re in.
What Designing AI for the Organization Actually Costs
None of this is an argument against AI. It’s an argument against implementing a cost reduction initiative and calling it customer service. It’s the AI-era version of, “Our intentions are honorable. We desire above all things to help.”
But make no mistake: Ease for organizations is not the same thing as ease for humans. A chatbot that deflects inbound contacts maximizes ease for the organization. Maybe the chatbot makes customers happy. But maybe it infuriates customers who want a customer service agent who can actually relate to their problem and deliver authentic empathy. A human.
This failure mode has a pattern that shows up the same way, again and again, across industries.
As organizations invested heavily in AI automation in 2024 and 2025, the volume metrics looked fine. Resolution rates, tickets handled, and time to first response all improved. (The Klarna AI press release that broke the internet says it all.) But the experience was deteriorating underneath those numbers, with customers reporting repeat explanations, time wasted, and, ultimately, declining trust.
Here’s what those customer-reported metrics actually feel like:
Bad AI customer support asks questions the system should already know. We get nudged down tidy little decision trees that bear no resemblance to our actual problems. We have to repeat ourselves because the first interaction didn’t really “count.” Bots speak like over-trained interns trying to sound upbeat while we all lose our goddamn minds.
The Qualtrics survey findings reflect this lived experience:
Half of respondents said their number one concern about AI in customer service is not being able to reach a human. Nearly one in five people who have used AI for customer service said they saw no value from it whatsoever. In fact, people rank AI-powered customer service among the worst applications of AI when it comes to convenience, time savings, and usefulness. And the only AI application people find less useful than AI customer service is using AI to build other AI. (Wasn’t customer service supposed to be the easy win?)
At Bodine & Co. we call this The AI ROI Mirage.
Organizations optimize for their own efficiency metrics and measure from the inside out, optimizing for what they can tally up on a dashboard — tickets handled, agents replaced, cost per interaction. But what looks like savings isn’t, because the human cost never appeared on the spreadsheet. The organization delivers an experience customers perceive as hostile, indifferent, or simply useless. So we all lose our minds and the dignity of being seen as a whole person rather than a support ticket.
Back to the ships
The aliens from “To Serve Man” never actually lie, and humans find out that their devices work exactly as promised. Meanwhile, the humans get answers that sound oh-so-right: “Just look at all the amazing stuff we’re giving you. Trust us.”
So they board the ships.
Then, just as the main character walks up the ramp for his intergalactic flight, his assistant runs across the tarmac, screaming, “Mr. Chambers! Don’t get on that ship!”
She’s finally translated the rest of the book the aliens had left behind.
“To Serve Man” is a cookbook.
It took them a while, but the humans figured it out.
Your customers are figuring it out, too. They know your AI was designed to serve your organization’s appetite for efficiency, not their need for help. They’re not confused about this. And they’re running out of reasons to keep boarding.
They’re also not filing complaints or asking to speak to a manager. They’re just stepping off the tarmac and not coming back.
So before the chatbot, before the automation roadmap, before the customer backlash — ask this question:
Who is this designed to serve?
***
P.S. One major difference about our situation is that, of course, there are no aliens. In our version of the story, we’re the ones making the promises about magic tech. We’re doing this to ourselves. So if you’re an AI decision-maker, be a lamb and side with humanity.
