How IKEA turned a €13 million chatbot into a €1.3 billion business

In 2021, Ingka Group, the main operator of IKEA stores, launched a chatbot called Billie. Its objective was typical of a conversational assistant: to answer routine customer inquiries, such as product availability, delivery times, or order status.

As is typical for the use case, Billie’s launch freed up call center teams from repetitive tasks. Between 2021 and 2023, Billie handled 3.2 million interactions and resolved approximately 47% of those inquiries, resulting in cost savings of €13 million.

On the horizon loomed an uncomfortable consequence: 8,500 people in call centers whose work, for the most part, was being done by machines. The direct question was clear: what to do with a workforce that was suddenly difficult to justify? It was a matter of costs, and the answer would have been predictable.

But IKEA did not ask itself that question.

Instead of trying to cut costs by reducing that team, the company looked in a different direction. They focused on where AI fell short.

While Billie answered 47% of the queries, that left 53% that it couldn’t resolve. After analyzing those requests, IKEA discovered that it wasn’t a matter of optimization that would be solved with the next version of the model. That’s because customers weren’t calling just to find out whether a sofa was in stock. They wanted to know whether it would look good in their living room, and they expected a person on the other end to advise them.

That changed everything. IKEA repurposed its 8,500 call center employees as remote interior design consultants. What had been a limited-reach service was scaled up to become a remote sales channel operated by thousands of human consultants via video and telephone.

In other words, the cost center became a revenue stream that generated €1.3 billion in its first full year, 3.3% of IKEA’s revenue. The company told Reuters that it aims to increase this to 10% by 2028. Four years later, the experiment is now a well-established strategy.

A new perspective on what AI reveals

Most AI initiatives focus on the AI ​​itself. In the case of a chatbot, organizations pursue metrics such as how many queries it handles, how much the cost per interaction decreases, or how many hours it frees up. These are the figures requested by management, provided by the vendor, and included in the reporting.

But there’s another perspective that goes in the opposite direction: focusing on what AI reveals that was previously unseen. In IKEA’s case, it was a need that its customers had had for decades and one that the company had never been able to identify as a business opportunity: an interior design consultant .

This situation isn’t limited to the furniture sector. For example, a legal team might encounter business inquiries they never had time to address, and a technical service that automates first-level support might uncover underlying problems that no one was paying attention to. Every sector has its own specific cases.

The difference between the two approaches translates into figures. Looking inward for AI efficiencies saved IKEA €13 million. Looking outward for new opportunities generated €1.3 billion in revenue. One hundred times more, to the point that the company barely mentions the initial savings.

AI solves known questions, but the real business value lies in the ones it opens up.

But identifying the opportunity is only half the battle. The other half is deciding what to do with it, and that’s where almost everyone backs out.

The immediate solution, the one no one discusses in a committee, was to cut a workforce that AI had rendered obsolete. Who, instead, decides to redefine 8,500 people and invest in a service that barely existed? It’s a risky strategic decision, one that goes against the grain — specifically against the direct savings shown in the reports.

Here’s the key: IKEA went beyond an AI project. It created a new function within the business. It didn’t just relocate people for the sake of relocating them. It identified an unmet need and built, on top of its existing in-house capabilities, a service capable of addressing it.

This way of thinking moves beyond the classic machine-vs.-person debate — i.e., how many jobs AI replaces or destroys — to enter another: what new opportunities AI reveals, and what we decide to build with it.

AI opens a new path for the CIO

It’s easy to think it was a visionary who discovered an opportunity. In this case, there wasn’t one. What there was, however, was a structure where the question could arise. And the key wasn’t that they had an AI task force (something common in large companies) but something almost no one has: the head of technology and the head of people sitting at the same table, not on two separate committees.

Creating these conditions is within the CIO’s reach. This case illustrates how AI presents a decision for the CIO, one that many are unaware of.

Two paths open up. One is to do what’s expected: execute, deploy the AI, measure its performance, report the savings, and move on to the next use case. It’s a respectable, and predictable, approach. The other is to open a different conversation: what does each deployment reveal, what unmet needs emerge when the tool takes over operations, what skills lie dormant in employees because no one has asked them to develop them? You can start by raising these questions in the next meeting; there’s no need to start by creating a new committee.

The first approach treats AI as a technical project. The second treats it as what it is: a business initiative that passes through the CIO’s hands before anyone else’s.

What’s at stake isn’t just the outcome of the initiative. It’s what kind of CIO emerges from each path. AI is rewriting the CIO’s role, whether they like it or not. The real choice is whether they sit in the pilot’s seat or the passenger seat.

It’s not about the CIO coming up with the next billion-dollar service, but about creating the space where that opportunity can emerge and reach those who recognize it. This involves making decisions that don’t generate headlines, such as bringing technology and business representatives together to compare their perspectives on AI — something that takes time and goes against rushing.

This demands something of the CIO that the role didn’t prepare them for: to make decisions with a different logic. Not the logic of fitting into the reporting framework, but the logic of seeking out opportunities that aren’t guaranteed. No one is going to force them to look beyond AI. That’s why almost no one does. But that’s precisely the space that AI opens up, and where they can define their path and gain new influence within the organization. AI isn’t taking their job away; it’s offering them a bigger one.