Customer Service Quality Under Attack

In the era of AI, one of the biggest losers has been the quality of customer service.

Not because AI is bad. Not because chatbots are useless. Not because automation cannot improve the customer experience. It absolutely can. The problem is simpler and uglier: many companies are using AI not to improve customer service, but to attack it, shrink it, hide it, and sometimes replace it with the cheapest possible imitation of support.

Customer service was always the ugly duckling of most companies.

It does not directly add revenue. It is expensive. It requires significant headcount. The workforce in customer service is usually different from the rest of the business: often younger, often less experienced, often looking at the job as a transition rather than a long-term career. Customer service is not sexy like marketing. It is not loud like legal or compliance. It does not usually have the glamour of product, the status of strategy, or the political weight of finance.

For all these reasons, and a few more, it is often treated as a wasteland where companies burn money without much benefit.

This could not be further from reality.

Over 80% of the services I cancelled in the last 24 months, including banks and insurance products, were cancelled because of poor customer service. Not because the product was useless. Not because the price was impossible. Not because I woke up one morning with a sudden desire to reorganise my direct debits. Poor customer service pushed me out.

I was a big advocate of my previous bank. I had been a client for more than five years. I had several products there. I recommended it to friends. Then one day I needed a chargeback.

What I got was a terrible chatbot and awful customer service.

Two months later, all my products, investments, and direct debits had been moved to another bank.

That is the part many executives still pretend not to understand. Yes, acquiring new clients is not the main strength of customer service. But losing them? That is where bad customer service truly shines.

You provide poor support, you lose clients. It is that simple.

The customer may not leave immediately. He may not send a dramatic farewell letter. He may not fill out your satisfaction survey. He may not write an angry LinkedIn post. He may simply stop trusting you. Then, when a competitor offers something “good enough,” he leaves. Quietly. Efficiently. Permanently.

And the company looks at a dashboard six months later wondering why retention is getting worse.

This is not an argument against chatbots or automated services. Quite the opposite.

A good chatbot can be a real quality improvement. Companies can save money and, at the same time, provide faster and better support. A chatbot that works 24/7, answers clearly, solves basic issues, routes complex cases properly, and gives customers immediate help is excellent. In many cases, it is better than waiting in a queue for an overworked agent who has to handle five different systems and three different policies at the same time.

But that is the hard way.

The hard way requires planning. It requires strategy. It requires proper prioritisation. It requires understanding why customers contact support in the first place. It requires knowing which contacts can be automated, which ones should never be automated, and which ones need a human being after the first failed attempt.

It requires the company to care about the outcome, not just the cost.

The easy way is very different.

The easy way is to say: “Let’s cut cost per contact.”

That sentence sounds professional. It sounds operational. It sounds like something a consultant would write on slide 14 of a transformation deck. But in practice, it often means this: make it harder for customers to reach support.

Remove the email address. Hide the phone number. Push everything into a chatbot. Build help-centre articles that answer every question except the real one. Close chats too quickly. Force customers to repeat the same information. Limit complaint forms to a few characters. Treat every human contact as a failure of deflection rather than an opportunity to solve a problem.

Then celebrate because volume went down.

Of course volume went down.

If you completely prevent customers from contacting you, you can deflect 100% of the volume. Congratulations. You have won the spreadsheet.

But you may also deflect the willingness of your customers to do business with you.

A customer who gives up is not a satisfied customer. A customer who cannot reach you is not a solved case. A customer trapped inside a chatbot loop is not proof of digital transformation. It is proof that the company has confused obstruction with efficiency.

That is what is happening now.

Customer service is not dead. It is being attacked by a very specific mentality: the idea that support is only a cost, that every contact is a problem, that every agent is an expense, and that every customer who needs help is somehow damaging the business model.

AI did not create that mentality. It only gave it better tools.

Before AI, bad customer service was slower, more visible, and harder to disguise. Now, it comes wrapped in modern language. “Automation.” “Self-service.” “Digital journey.” “AI-powered experience.” “Contact deflection.” The vocabulary is new. The instinct is old.

Cut the cost. Hide the pain. Hope the customer stays.

But customers are not stupid. They know the difference between self-service and no service. They know when automation is helping them and when it is being used to block them. They know when a chatbot is there to solve a problem and when it is there to exhaust them.

And once that trust is gone, it is very difficult to rebuild.

The future of customer service should be quality.

Aim first for quality, then for cost savings. You can do both. In fact, the best companies will do both. They will use AI to remove repetitive work, reduce waiting times, support agents, improve routing, summarise cases, detect urgency, and solve simple problems instantly. They will not use AI as a wall between the customer and the company.

That is the difference.

Good automation serves the customer faster.

Bad automation makes the customer disappear.

Customer service does not need to be protected from technology. It needs to be protected from lazy management using technology as an excuse to do what it already wanted to do: spend less, listen less, care less.

The companies that understand this will win. Not because they will have the most expensive support departments, but because they will understand what customer service really is.

It is not a wasteland.

It is not a charity project.

It is not just a cost centre.

It is the place where a company either keeps its promises or exposes them as marketing.

And in the era of AI, that distinction will matter more than ever.