IKEA reskilled 8,500 employees into interior designers: how did they get AI implementation so right?

There are two questions a business can ask when introducing AI to the business – be they an established enterprise or a startup. The first is: how do we cut costs with AI? The second is: what can our people do now that AI handles the routine? The question asked determines everything that follows.

Right now, across the global economy, two radically different AI strategies are playing out simultaneously. They start from the same technology, but they produce completely opposite outcomes. The difference between them is not technical capability, investment scale, or market position. It’s mindset.

Strategy one: cut

Oracle and dozens of others are cutting tens of thousands of people to fund AI investments. The logic behind these decisions is straightforward – if AI can do the work, we need fewer humans. Deploy the technology, reduce the headcount and report the efficiency gains.

Nvidia CEO Jensen Huang – the man whose company powers most of the world’s AI infrastructure – has been unusually blunt about this approach. “The narrative that connects AI to job loss, for many of the CEOs that are doing it – it is just too lazy,” he told Channel News Asia earlier this year. His point was sharp: AI has barely arrived, and companies are already claiming it justifies layoffs that in reality have nothing to do with the technology.

Huang’s deeper argument is the one that matters most for leaders thinking about AI strategy. He told CNN: “If the world runs out of ideas, then productivity gains translates to job loss.” In other words, companies with ideas suffer from not having enough people to execute them. AI should give them more capacity to pursue those ideas. But if the first instinct is to use AI to reduce the workforce, the company is revealing that it doesn’t have enough ideas to justify the people it already has. “In time,” Huang said “you’re just going to be a company that shrinks”.

This is the attrition spiral many of us have been talking about – but articulated by the CEO of a four-trillion-dollar company. Cut headcount to fund AI. Lose another tranche through voluntary attrition from the survivors. Recruit replacements at premium cost and net out to zero or worse. All the while, the ideas that would have justified the original workforce go unexecuted because nobody is left to pursue them.

There is a phycological dimension to this that leaders rarely discuss. When a company announces AI-driven headcount reductions, every remaining employee receives the same message; your value is conditional on being cheaper than a machine. The employees who stay don’t stay with enthusiasm – they stay with vigilance. And vigilant employees don’t innovate, take risks or bring discretionary effort; critical in the startup world. They protect themselves. The company gets compliance, but loses the creative energy that generates the ideas Huang is talking about.

Strategy two: grow

Then there are companies like IKEA.

IKEA deployed an AI chatbot called Billie to handle first-level customer service inquiries such as order tracking, product information, and returns. Billie resolved approximately 47% of all inbound conversations without human intervention. By any standard efficiency metric, the deployment was a success. According to the standard playbook, the next step would have been an obvious one: reduce customer service headcount by 47%.

IKEA didn’t make that choice, instead, they asked the question that most companies skip entirely. What are the conversations AI can’t handle, and what do they tell us about what our customers actually need?

The answer was revealing. The 53% of conversations that Billie couldn’t resolve weren’t failures of the technology. They were a demand signal. Customers were reaching out for help designing their homes. They wanted advice on room layouts, furniture combinations, special planning hints, and tips. These were consultative, creative, deeply human conversations that required taste, empathy, and contextual judgment, which is exactly the kind of work AI cannot do.

As a result, IKEA reskilled 8,500 customer service employees as interior design consultants. The same people, now empowered with new skills in new roles. The AI handled the transactional work, and the humans handled the work that required human expertise. The result: €1.3 billion in new revenue in the first year. That’s 3.3% of total group revenue, from a business line that barely existed before. IKEA’s not stopping there with a target is to grow it to 10% by 2028.

See Also

Building a business that puts people first and still wins big

The contrast with the “cut” strategy laid out earlier is worth sitting with. Every one of those 8,500 IKEA employees received the opposite message from the employees of companies using AI to reduce headcount. Rather than “Your value is conditional on being cheaper than a machine”, they heard “We believe you can do something more valuable than you are currently, and we are going to invest in helping you get there.” That is the foundation of a deep organisational loyalty.

Why this is a joy mapping story

IKEA didn’t use the language of “Joy Mapping” – the practice of understanding what energises people at work, where they need more time, and what they’d rather be doing with it. But they did it by instinct. They asked their customer service teams what the AI couldn’t handle, they listened to the answer, and they redesigned roles around what they heard. The AI was deployed to serve the humans, not replace them, and the humans were elevated, not eliminated. For startups, this demonstrates why understanding the ‘why’ of AI implementation matters before the technology arrives. Consider what IKEA did through the lens of three questions:

• “What part of your job drains you?”: Repetitive, transactional customer service queries like order status updates, returns and product availability. This is the work that Billie now handles
• “Where do you need more time?”: Customer conversations that require real expertise, such as design advice, spatial planning and creative problem-solving. The conversations that customer service agents were having in fragments between the transactional calls, never with enough time to do them properly
• “What would you do with that time if you had it?”: Interior design consulting to help customers create homes they love. Work that uses human judgment, empathy, and creativity. Work that generates joy and, as it turns out, €1.3 billion in new revenue.

For IKEA, the AI didn’t eliminate 8,500 jobs. It revealed 8,500 better ones. And the revenue it generated didn’t come from cutting costs, it came from the ideas that only became visible once the routine burden was removed.

The lesson for new business owners

Jensen Huang is right: a company that uses AI to shrink its workforce is a company that has run out of ideas. But the corollary is just as important: a company that uses AI to free its workforce to pursue better ideas is a company that has just discovered how many ideas it actually has.

The difference between these two outcomes is a single question that new businesses owners need to ask before deployment rather than after. Not “what can AI do?” but “what can our people do now that AI handles the routine?” That is the true Joy Map question. It’s the IKEA question. And it’s the question that separates the companies that will grow from the ones that will, in Huang’s words, shrink in one place. AI doesn’t decide what happens to your people. You do.