Bill Lewis Linacre Capital

Writing · Customer Relationship · September 2026

You Engineered Away the One Thing You Called Priceless

You had an asset you described, in writing, as the most valuable thing your company owned. The customer relationship. It is in your strategy. It is in your annual report. It is in the language you use about yourselves when you want to sound serious.

Then you took it apart on purpose, and called it efficiency.

Not by accident. Not through neglect. Somebody built a business case. Somebody costed out the person the customer used to speak to — the person they had, most likely, built up a relationship with — put a saving next to the empty chair, and presented it to you as progress. And you approved it, and felt shrewd.

This is about what you actually removed when you did that — and why the person who signed it off may come to regret it more than anyone.

What the human was for

Think about the moments a customer reaches for a person. Not to check a balance. Not to track a parcel. Those are the easy moments, and we will come back to them.

They reach for a person when something has gone wrong. The bill that makes no sense. The claim that was refused. The account locked for a reason no screen will explain. The circumstance the form did not anticipate. The decision they want to challenge.

Those moments have one thing in common. They were the moments a customer could hold your company to account, in real time, there and then. The customer spoke to a named person who had to answer. And that person could do something no system can. They could listen. They could decide your case was different. They could say, “I understand. Let me help. Let me fix it for you” — and they had the authority to mean it.

That is what you engineered away. Not a cost line. The moment when a customer still had agency — the single point at which they could push back and be heard.

You removed the customer’s power and kept your own

Look at what is left. A chatbot that returns to its script. A phone line that loops back to the menu. A portal with no field for the actual problem. An app that can do only what it was configured to do.

Every one of those channels is one you control. You decide what the menu offers. You decide what the bot will discuss. You decide what the form will accept, and what it will simply refuse to hear.

The human conversation was the only channel the customer controlled. It was where they could refuse to be routed to an empty hole. They could insist. They could demand you escalate. They could ask the question no script anticipated. It was their leverage — and you removed it, while telling them, and telling yourselves, that you were doing them a favour.

That is a very comfortable place for an institution to stand. Whether it is a company, a bank, a telco, a water utility, a local authority, a central government department, or your local GP surgery — it is comfortable for them, and useless for the customer. And the customer who finds themselves on the wrong side of it walks away, whenever and wherever they still can.

A system never has to answer “why”

Here is why that matters more than a complaint about service.

A customer can ask a person why. Why was I charged this. Why was my claim turned down. Why can I not see my own record. Why has this taken six weeks. And a human on the line has to engage with the question — and sometimes has to concede the answer is indefensible.

A system has no such duty. It will tell the customer what — what the balance is, what the policy says, what the next step is. It cannot be made to answer why, because it was never built to. So when you automated the front door, you did not just cut cost. You made your own company, at the exact moments a customer most needs to challenge it, almost impossible to challenge.

And now the concession, because it matters

None of this is an argument against automation. Most of it is genuinely good, and, in many cases, no customer wants the human back. Renewing online in two minutes. Checking a balance at midnight. Tracking an order without waiting on hold. Better than what it replaced, cheaper for you, more convenient for them — all at once.

The problem was never that you automated. It is where you automated. You applied it to the easy moments, which was sensible. Then you applied it to the hard ones — the moments that needed judgement, an exception, an answer — because those cost the most. And those were the only ones that were ever holding the relationship together.

Now the part that should trouble you most — because it is not about sentiment at all

Set the customer aside for a moment. Set aside loyalty, trust, all of it. Look only at your own costs, which is what the efficiency case claimed to be about.

You removed the human to save money. But the customer’s problem did not leave when the human did. It is still there. So the customer comes back. The chatbot could not help, so they call. The call did not resolve it, so they call again. They email. They complain. They complain a second time. They find another number. Every one of those contacts is one your business now has to handle and pay for — and not one of them would exist if the problem had been solved the first time, by a person, in a single conversation.

You did not remove the work. You multiplied it. You took one expensive human interaction and turned it into five cheap ones that do not add up to a resolution — and you are paying for all five.

Now the scale. When organisations actually measure this, they find that somewhere between 40 and 60 per cent of everything their contact operation handles is not real demand at all. It is this: contacts that exist only because the customer was not sorted the first time. In parts of the public sector, figures as high as 80 per cent have been reported. Read that in your own terms. On these numbers, up to half of what your service operation costs you is self-inflicted — work you manufactured by trying not to pay for the human who would have prevented it.

This is not a new idea, and it is not a fringe one. It has been understood for decades, and it has a name: failure demand — the work you are forced to do, and pay for, purely because you failed the customer the first time. The term was coined by the British occupational psychologist John Seddon.

So look again at the empty chair. You costed that person as an expense and removed them. But that person was the cheapest thing in the entire system — because a human who resolves a problem once, properly, prevents the flood of repeat contacts that a machine failing to resolve it generates. They were not a cost. They were load-bearing. You knocked out the cheapest, most valuable part of the structure, and the cost did not disappear. It moved to a part of the business where you were not looking — and it grew on the way.

The asset you said was everything

A relationship is not made in the easy moments. It is made in the hard ones — when something has gone wrong and a person on your side puts it right. That is where loyalty is actually earned. And that is the precise moment you removed, because it was the moment that cost the most and asked the most of you.

You did not decide the relationship no longer mattered. You decided it was “too expensive”. And you dismantled the one place where problems were solved and relationships built, without ever quite admitting to yourselves that was what you were doing.

So here is the account, in the only terms the decision was ever justified in.

You set out to serve the shareholder, and you made the shareholder poorer — because the saving on the front door was smaller than the cost you created behind it. You set out to protect the most valuable asset you own, and you destroyed the one place it was built. You did both of those things with a single decision, and you were more certain of that decision than almost any other you made that year.

You called it priceless. Then you engineered it away — and you were wrong about the one thing you were most sure of.

About the author

Linacre Capital works with boards on exactly this kind of decision: the ones that arrive dressed as efficiencies, carry no warning label, and turn out to have moved the risk rather than removed it. The hardest problems a board faces are rarely the ones that announce themselves. They are the ones already approved, already filed, already counted as a win.

Bill Lewis is Founding Partner of Linacre Capital Partners. He provides independent counsel to chairs, CEOs and founders on their highest-stakes decisions — bill@linacre.net.
Visit Linacre Capital.