The Case for Keeping Humans in Your Contact Centre
By Mike Powrie, CEO, NeonNow
In 2024, Klarna told the world its AI had replaced 700 customer service agents. By 2025, the company was rehiring them.
That story has been told a hundred different ways. Cautionary tale. Cost-cutting backfire. AI hype meets reality. Most of those takes are missing the point.
The Klarna reversal is not a story about AI being bad. It is a story about what happens when you confuse what AI is good at with what your customers actually want. And it is a story every contact centre leader should be paying close attention to right now, because the same mistake is being made across the industry every week.
Klarna's CEO, Sebastian Siemiatkowski, said it himself when the company started bringing humans back. The all-AI approach produced lower quality customer service than the people it replaced. That is a remarkable thing for a tech-forward CEO to admit publicly, and it is the most useful sentence in the entire conversation about AI in contact centres right now.
Because here is the thing nobody selling AI wants you to think about: your customers know.
What your customers actually want
Metrigy's 2025-26 Customer Experience Optimization study asked over 500 consumers a simple question. Would you rather speak to a human or an AI agent? 84.7 percent said human. Even when researchers reassured them that the AI could resolve their issue, 80.1 percent still wanted a person.
That number should give every CX leader pause. Not because it means AI has no place in your contact centre. It absolutely does. But because it tells you something important about the gap between what is technically possible and what customers actually want.
Other research lands in the same place. A SurveyMonkey study found that 79 percent of consumers strongly prefer interacting with a human over an AI agent. A separate piece of research from Kinsta showed 41 percent of consumers feel customer service has actively got worse because of AI. 63 percent do not believe AI could ever replace humans in customer service at all.
There is nuance in this data, of course. Consumers happily use AI for some things. Tracking deliveries. Confirming appointments. Getting routed to the right team. Quick, transactional, low-stakes interactions where speed beats nuance. But for anything that matters, anything emotional, anything complex, anything requiring judgement, people want to talk to a person. They always have. They probably always will.
That is not a problem your AI strategy needs to solve. It is a feature of being human.
The contrarian case nobody is making
Here is the part of this conversation that I think gets missed.
For years, the dominant industry narrative has been that human agents are an expensive constraint. A cost line to be reduced. A workforce problem to be solved with technology. The framing was always: how do we do more with fewer people?
I think that framing is about to flip.
If the data is right, and customer preference is shifting against over-automation, then organisations that have invested in keeping smart, capable human agents in their contact centres are not behind. They are ahead. They have the capacity to deliver the experience customers actually want, while their competitors are scrambling to rehire after a botched AI rollout.
Gartner predicts that more than 40 percent of agentic AI projects will be cancelled by the end of 2027. Not because the technology does not work. Because organisations are deploying it badly, measuring it wrong, and discovering too late that the savings on paper do not match the costs in practice.
When those projects get cancelled, what do those organisations do? They go back to humans. Just like Klarna did. And the cost of unwinding a failed AI replacement programme, of recruiting, of training, of rebuilding the culture you stripped out, is enormous. It rarely gets factored into the original business case.
So here is the contrarian view. In a market where almost every CX vendor is racing to sell you automation, the smartest move might be to think very carefully before automating the wrong things. Your competitors are running into walls right now. The opportunity is not to follow them.
What "human plus AI done right" actually looks like
None of this is an argument against AI in your contact centre. Let me be very clear about that. The story is not humans versus AI. It is humans plus AI, deployed thoughtfully, with intent, in the places it actually creates value.
What I see working well across our customers, and what the data points to as the consensus best practice, is something like this:
AI handles the high-volume, repetitive, transactional interactions where speed beats empathy. Booking confirmations. Status updates. Simple lookups. Routing the caller to the right team. These are the things customers do not want to wait on hold for. AI does them faster than humans, often better, and at a fraction of the cost.
Humans handle everything else. Complaints. Emotional conversations. Anything involving judgement or compassion. Anything where the customer is frustrated or vulnerable. Anything that requires real conversation, not just an answer.
And here is the part that matters most. The AI needs to make the humans better at their jobs. Not replace them. Not work around them. Make them better.
That means giving agents real-time prompts during difficult calls. Surfacing the right information at the right moment. Detecting when a customer is becoming frustrated so a supervisor can step in. Summarising calls automatically so agents do not waste twenty minutes after every interaction on wrap-up. Spotting patterns across thousands of conversations that no individual manager could ever see manually.
These are not future-state features. They are real, deployed capabilities running on enterprise contact centres today. The features list inside Amazon Connect is a good reference point if you want to see what is now considered standard at the platform layer. Real-time agent assist. Automated call summaries. AI-powered call analytics. The direction of travel is consistent across the platforms that matter.
This is exactly what we built NeonNow IQ to do. It is the AI intelligence layer that runs across the entire NeonNow platform. It is not there to replace your agents. It is there to make every conversation they have go a little better than the last one. To take the stuff that drains them and the stuff they hate, and let them focus on the work that only humans can do.
That work, by the way, is also the work that customers value most. The connection. The empathy. The sense that someone on the other end of the line actually cares whether you walk away happy. You cannot fake that with a chatbot, no matter how impressive the language model is.
Why this matters now
If you are sitting in a CX leadership role right now, you are probably being asked the same question I am asked every week. What is your AI strategy?
It is a fair question. AI is going to be a critical part of how every contact centre operates from here on out. The technology is real, the value is real, and the organisations that figure out how to deploy it well are going to have a meaningful competitive advantage over the ones that do not.
But the question behind the question matters too. What is your human strategy?
Because here is what I would bet on. The contact centres that win the next five years are not going to be the ones with the most AI. They are going to be the ones who got the balance right. AI for the volume. Humans for the moments. A platform underneath that brings both together properly, on infrastructure that does not break when things scale.
That last part matters more than it sounds, by the way. We built NeonNow on AWS and Amazon Connect for exactly this reason. Enterprise contact centres need an infrastructure foundation that just works. Security, scalability, uptime, compliance. The stuff that does not show up in marketing presentations but determines whether your platform survives contact with reality. Amazon Connect is the most mature cloud contact centre platform on the market, and AWS gives us the confidence that what our customers build is going to hold up.
So if you are reviewing your AI strategy, here is the question I would ask before any other. Are you using AI to make your humans better, or to replace them? Because Klarna can tell you exactly what happens when you choose the second option.
The opportunity, simply put
Your customers want to talk to your people. That is not a problem. It is a gift. It is the thing that competitors with thousands of cancelled AI projects will spend the next decade trying to rebuild.
Keep your humans. Use AI to make them brilliant. That is the strategy.
It is not flashy. It will not get you a press release with a stat about replacing 700 agents. But it will keep your customers coming back, and it will give your team something to be proud of building.
In a market full of vendors selling you a future without people, choosing to invest in your people might be the most radical move you can make.
If you want to dig into how the underlying platforms are evolving, the AWS Contact Center blog is one of the better places to follow what is genuinely being built versus what is just being marketed. I read it most weeks.
Want to see what AI plus humans actually looks like in practice?
NeonNow IQ runs underneath every conversation in your contact centre. It supports agents in real time, surfaces insight automatically, and handles the volume so your team can focus on the moments that matter. All on a platform built on AWS for the reliability and security your business needs.
Book a demo and we will show you exactly how it works.
Or if you would rather have a conversation first, contact us and let's talk through what good looks like for your contact centre.
.jpg)
.svg)
