How IKEA avoided layoffs, and made a billion

IKEA’s new AI-powered customer service bot now handles 57% of customer conversations without a human agent, covering order tracking, product queries, and general support.

This is not a plan. It is already running.

What happens when routine is automated?

When IKEA deployed its AI chatbot named Billy, the company could have done what seems to be the trend across the business world: lay off customer service staff and cut costs. That is exactly what Klarna did. The Swedish fintech company cut 700 customer service workers to make way for AI. Customer satisfaction fell, complaints rose, and CEO Sebastian Siemiatkowski publicly admitted: “We went too far.” Klarna is now trying to hire people back.

IKEA took a different path. Instead of layoffs, management analysed what was actually happening in the remaining 43% of customer interactions. Those were not transactions. They were advisory: people wanting help designing their home or office space.

IKEA retrained its customer service staff as interior design consultants. In the first year, the new consultancy business generated nearly 1 billion euros in additional revenue.

Is AI already visible in employment statistics?

Data published by the BLS (United States Bureau of Labor Statistics) in May 2026 provide the first official statistical evidence that AI is causing targeted occupational displacement: 18 professions classified as likely to be affected by AI saw a collective employment decline of 0.2%. Among sales representatives, the drop was 4.8%.

At the same time, in Finland, the Prime Minister’s Office (valtioneuvosto) reported in spring 2026 that for the first time in the history of its Kansalaispulssi survey, which has been conducted since spring 2020, employment topped the list of priorities Finns want the government to address. Over half of respondents named it as one of their three most important areas. Young people’s confidence in the future has fallen from two thirds to 50% in a single year.

Together, these two data points are hard to dismiss: the AI transition is real, already under way, and unevenly distributed.

The IKEA model is not self-evident

It is worth being honest: IKEA is a large company with the resources for significant retraining investment. Most smaller employers and public sector organisations do not have the same room for manoeuvre.

That is precisely why this is a political question, not just a management one.

A useful point of comparison is our dear neighbour Estonia. Prime Minister Kristen Michal has set a target of doubling the country’s GDP by 2035 through AI. In Finland, equivalent political ownership of this agenda does not yet exist.

Invest in retraining before the layoffs

The IKEA case shows when investment works best: before automation takes the job, not after. After-the-fact support is important, but it is more expensive and slower than proactive training.

For Finnish policy, this means three things:

• Fund continuous learning: in the AI era, professional renewal cannot rely on degree programmes alone. Modular, short-cycle upskilling pathways are needed, accessible to both businesses and individuals.
• Support for SMEs: large companies can implement the IKEA model with their own resources. Finland’s business structure is SME-driven. Mechanisms are needed to enable the same transition in smaller organisations.
• Political ownership: the AI transition is not one policy area among many, but a complete rethinking of how the public sector operates. It requires consistent political leadership, not pilot projects alone.

AI does not replace people, but it forces a reckoning with where value is actually created. And that demands a political response.

Researchers Jari Hakanen and Janne Kaltiainen of the Finnish Institute of Occupational Health (Työterveyslaitos) wrote in April 2026 in Helsingin Sanomat that hope does not mean passive waiting. It means strong agency and the capacity to find paths forward even in the middle of change.

IKEA’s example says the same thing in the language of management: AI does not decide what happens to people. Decision-makers do.

What do you think, does Finland have what it takes to implement the IKEA model at scale?