Arthur Wickstead was peeling the adhesive backing from a “Priority” sticker when the phone rang for the that morning. He was , his knuckles were permanently stained with the faint violet ink of the carbon copies he refused to abandon, and he was currently staring at a return request that made absolutely no sense on paper but made perfect sense in the world of human error.
A contractor in northern Maine had ordered a multi-zone condenser, but through a series of logistical hiccups involving a diverted freight truck and a blizzard that had effectively shut down the I-95, the unit had arrived late. By the time it landed on the gravel driveway, the homeowner had already panicked and bought a different system from a local hardware store.
Technically, the return window had closed. Technically, the contractor was stuck with a $2,140 piece of equipment he couldn’t use. Arthur didn’t look at the clock. He didn’t look at a dashboard of Key Performance Indicators. He looked at the date on the bill of lading, noted the weather report from the week of the 11th, and realized that expecting a freight carrier to navigate a whiteout was a mathematical absurdity.
He picked up his pen, scrawled “Approved – Act of God” across the top of the form, and authorized the refund. He was an “inconsistency” in the company’s data set. He was a variable that could not be predicted by a spreadsheet.
01
The Automation of Indifference
Six months after Arthur was offered a generous early retirement package, the company replaced his desk with a software suite. They called it a “Rules Engine.” It was heralded as a triumph of modern logistics, a way to ensure that every customer was treated exactly the same, regardless of who picked up the phone.
The board of directors loved the phrase “perfectly consistent.” They didn’t realize that in their quest to eliminate the erratic nature of human judgment, they were about to build a machine that was perfectly incapable of recognizing the truth.
rules_engine.exe
v4.2.0
[Error: Weather_Pattern_Not_Detected]
[Error: Human_Context_Value_Missing]
We have entered an era where the “Clerk” is seen as a liability. In the architectural blueprints of modern commerce, the person with the power to say “this is an exception” is viewed as a leak in the plumbing. We want systems that are frictionless, and humans, with our pesky emotions and our tendency to understand context, are the ultimate source of friction.
We cleared our browser caches in desperation, hoping the refresh button would reveal a hidden human element, but the logic gates remained shut.
The Binary Trap
The failure of the automated returns process is not a failure of technology, but a failure of philosophy. It is the belief that fairness is a derivative of consistency. If we treat everyone exactly the same, the logic goes, we are being fair. But this ignores the reality that not everyone arrives at the gate under the same conditions.
When a customer service portal denies a return because the “system” sees a date that is past a deadline, it isn’t being objective; it is being blind. It cannot see the funeral that kept the customer from the post office, or the defective flare nut that caused the refrigerant leak, or the fact that the shipping box was dropped by a driver who was on his of a shift.
Rigid, Clean, Simplified
Complex, Raw, Real
The old clerk understood that the “rules” were a map, but the “situation” was the territory. He knew that the map is often drawn in a warm office by people who have never had to drag a heat pump up a flight of stairs in the rain. When we automate that discretion away, we aren’t just making things faster; we are removing the safety valve that keeps a fair situation from becoming a bureaucratic nightmare.
The Edge Case is the Core Case
Consider the case of a modern HVAC installation. These are complex machines, sensitive to the environments in which they are placed. When a homeowner invests in a high-efficiency system, they aren’t just buying a box of metal and copper; they are buying the promise of climate control.
If that system arrives with a hairline fracture in the coil-one that isn’t visible until the system is pressurized-an automated return system might flag the return as “invalid” because the unit was “installed.” The software sees a binary state: Unused vs. Used. It does not have the capacity to understand that the defect only reveals itself through the act of use.
The real friction usually starts when the technical problem meets the logistical wall, which is why having access to professional support is often less about reading a manual and more about finding a professional who understands the physics of a failed installation and has the authority to rectify it. Without that bridge, the customer is left shouting into a void of “If/Then” statements that have no room for their specific reality.
The Tax of Efficiency
We often mistake automation for progress because it produces a clean data set. A computer never has a bad day. It never gets tired. It never lets its personal feelings about a rude caller influence its decision. These are all perceived as benefits. But a computer also never has a good day. It never feels the satisfaction of helping someone out of a bind. It never realizes that “Policy 4.2” was written for a different product line and shouldn’t apply to the current situation.
The unquantifiable cost of rigid uniformity in customer service environments.
The “Old Clerk” was a curator of common sense. He was the person who knew that if a long-time dealer who has spent with the company over asks for a one-day grace period on a return, you give it to him. Not because you are breaking the rules, but because you are protecting the relationship.
The automated system, however, sees the $80,000 dealer and the one-time buyer exactly the same. It sees them as account numbers with associated date stamps. In its quest for “fairness” through consistency, it alienates the very people who keep the business alive.
This transition from judgment to logic gates is a slow-motion erosion of trust. When a customer realizes they are arguing with a script, their blood pressure rises because they know that logic cannot be reasoned with. You cannot plead your case to an algorithm. You cannot explain the nuance of a shipping delay to a dropdown menu.
The “Denied” screen is a wall, not a conversation. And when we replace conversations with walls, we shouldn’t be surprised when customers stop trying to climb them and simply walk away to a competitor who still has a human on the other end of the line.
The irony is that the more “efficient” we make our returns and support systems, the more we increase the “hidden tax” of customer churn. We save money on the clerk’s salary, and we save money on the “erroneous” returns he might have approved, but we lose an unquantifiable amount of brand equity. We trade the violet ink of Arthur Wickstead’s pen for the red ink of a declining customer retention rate.
The Contractor’s Apprentice
I remember a specific instance where a customer had purchased a multi-zone mini split system. It was a complex setup, involving three indoor units and a single outdoor condenser. During the installation, one of the indoor units was accidentally damaged-not by a manufacturing defect, but by a contractor’s apprentice who had tripped over a ladder.
“The contractor, an honest man, called the company. He didn’t lie. He didn’t claim it arrived broken. He asked if there was any way he could get a discount on a replacement unit, given that he had just bought the entire $4,500 package.”
The automated system would have rejected this out of hand. “Customer-inflicted damage” is a hard “No” in any rules engine. But the human on the line-a person who understood that a gesture of goodwill would secure a lifetime of loyalty from a contractor who does a year-was able to make it happen. That wasn’t an “inconsistency.” It was an investment.
We are currently obsessed with the idea of “scaling” empathy. We try to build it into our UI/UX designs. We try to program it into our AI chatbots. We give them names like “HelperBot” and give them friendly avatars. But empathy cannot be scaled because empathy is a bespoke product. It is a one-to-one exchange of understanding.
When you try to automate it, it turns into something else-a performance of care that feels more insulting than a flat refusal. The loss of the “Old Clerk” is the loss of the grey area. We have moved into a world of high-contrast black and white, where you are either “In Policy” or “Out of Policy.”
But most of life happens in the grey. Most business happens in the grey. The grey is where the shipping delays happen, where the misunderstandings occur, and where the “Act of God” weather events reside. By eliminating the grey, we have made our businesses more “efficient” but significantly less human.
Hearing Voices
This is the central paradox of modern commerce: we have more tools than ever to “connect” with our customers, yet we are more disconnected from their actual experiences than we were ago.
Arthur Wickstead knew his customers’ voices. He knew which ones were reliable and which ones were prone to exaggeration. He had “Data,” but it was stored in his head and his heart, not on a server in Virginia. That data allowed him to be “inconsistent” in a way that was profoundly fair.
If we want to fix the broken state of customer support and returns, we have to stop treating “human judgment” as a bug to be fixed. We have to start seeing it as a feature that needs to be protected. We need systems that provide the data, but we need humans who provide the “Why.”
The goal should not be a “perfectly consistent” system. The goal should be a “perfectly fair” one. And fairness requires the one thing a computer will never have: the ability to look at a smudge on a bill of lading, hear the fatigue in a contractor’s voice, and decide that the rules are less important than the person standing on the other side of them.
As we continue to build more sophisticated engines for our businesses, we would do well to remember the violet ink on Arthur’s knuckles. We should remember that the most valuable thing he brought to the office every day wasn’t his ability to follow the manual, but his ability to know when to close it.
In a world of rigid logic, the most “advanced” technology we can offer our customers is the simple, inconsistent, and profoundly necessary gift of being heard by another human being. Without that, we aren’t building businesses; we are just building very expensive, very efficient ways to lose the people who matter most.