Ikea Retrains 8,500 Workers as AI Handles Routine Tasks

Five years before Ikea's bet took shape, call center workers like Melanie Lindell fielded routine questions—store hours, pet policies—from customers sitting in cubicles in an office park 65 kilometers north of Malmö, Sweden. In 2021, Ikea introduced Billie, an AI-powered customer service bot named after the company's ubiquitous Billy bookcase. The bot could absorb those repetitive queries with ease, assisting 47% of customers in its first two years; the figure has now reached 74%.

Instead of cutting the workforce, Ikea retrained approximately 8,500 call center employees over two years to handle more complex customer queries and work as interior design sales consultants. The training—five to six weeks for new hires—includes digital room-planning tools, communication techniques, and specialized product knowledge. Richard Richardson, who works at a remote-sales center in Sheffield, England, completed a two-week "kitchen school" and field visits to study Ikea's product lines in person. He learned to notice how wood grain aligned along joints and now advises customers on pairings like the Havsen sink with a drawer front to cover the waste pipe. More importantly, he learned to let customers speak before offering solutions—a technique that also creates opportunities for upselling.

The remote-sales centers—24 locations covering 31 countries where Ingka Group, Ikea's primary owner, operates—have become the company's fastest-growing sales channel. They generated $1.37 billion（約2200億円） in revenue last fiscal year, up from $1.17 billion（約1900億円） the prior year, with consistent 15%–20% annual growth. Last fiscal year, Ikea's workers helped 10 million customers through these centers. Customer happiness scores rose to 89% from 60% after Billie's rollout, suggesting that automation paired with strategic human upskilling improved rather than degraded the experience.

Madeleine Barr, a senior sales specialist who has designed kitchens at Ikea for more than a decade, exemplifies the value Ikea sees in human expertise. When a customer called with dimensions for a lopsided hexagonal countertop—"no two sides the same length"—with an electric stove in the middle, Barr worked through the geometry and returned with a solution: extending the overhang slightly would make the dimensions work. That judgment call, paired with Barr's ability to reassure and guide the customer toward a sale that might otherwise have fallen through, represents the kind of work Ikea believes machines cannot yet replicate. Barr unpacks physical slab samples at the Helsingborg center so advisors can touch real materials; she reports that customers want to know what surfaces feel like to the fingertips and whether knobs click satisfyingly—tactile and emotional information that bots cannot provide.

Ikea's approach stands in sharp relief to broader industry practice. An estimated 92 million jobs could be displaced by AI, related technologies, and demographic shifts by 2030, according to the World Economic Forum; a February 2026 Mercer survey found that 40% of workers fear AI will make their job obsolete, up from 28% in 2024. Ikea's private structure—both Ingka Group and Inter Ikea Group are privately held—insulated it from shareholder pressure to maximize cost-cutting through layoffs. While both companies laid off hundreds of corporate workers in spring 2026 to maintain affordability, none of the cuts affected remote-sales centers, nor were they attributed to AI. The company faces competitive pressure from Wayfair, Amazon, and design-forward retailers like Article, making every sale critical as Ikea's biggest retail arms reported revenue drops in fiscal 2025 despite higher customer visits and units sold.

Chief digital officer Parag Parekh did not rule out future automation of the design-advisory roles. He suggested any future layoffs would likely stem from macroeconomic factors rather than advancing AI capabilities, but the boundary between those categories remains uncertain. One stabilizing factor: despite offering free, partially automated design tools on its website, some customers continue to opt for consulting sessions with human advisors. Parekh also emphasized that home renovation is expensive and intricate—positioning a dishwasher so its open door does not obstruct passage between sink and stove is the sort of nuanced problem that some customers prefer to solve with human input. Whether future versions of Billie will respect that boundary remains the hanging question over Ikea's experiment.

Ikea's response to automation reveals a deliberate fork from the playbook most large companies follow when deploying AI. Rather than viewing automation as a way to eliminate labor costs, the company used AI to absorb routine, low-value tasks—basic customer service inquiries handled by Billie now account for 74% of bot-assisted interactions—and freed human workers to perform roles that command higher margins and drive customer loyalty. The shift was not spontaneous; it followed a crisis. When COVID-19 forced store closures, Ikea had to pivot rapidly to e-commerce, overwhelming its call centers with three times normal volume. This necessity created the opening for Billie, but the company's private structure (owned by Ingka Group and Inter Ikea Group) gave it freedom to experiment with reskilling rather than cost-cutting. Prasanna Tambe, a Wharton professor, frames this as leveraging "skill adjacencies"—using workers' existing domain expertise in customer interaction and product knowledge to move them into roles that satisfy customer wants rather than just needs.

The financial results suggest the bet is working. Remote-sales centers have grown 15%–20% annually and generated $1.37 billion（約2200億円） in revenue last fiscal year. Customer happiness scores rose to 89% from 60% after Billie's rollout. Yet the gamble carries real uncertainty. The company explicitly did not rule out future AI-driven layoffs; leadership suggested any cuts would stem from macroeconomic conditions rather than advancing AI, but that distinction may blur in practice. The hexagonal countertop anecdote—a problem "far too unruly for AI to solve"—illustrates both the current limits of automation (Billie cannot navigate irregular geometry or offer nuanced reassurance) and the potential fragility of this model: as AI capabilities improve, the boundary between automatable and human-only work will shift, and no guarantees exist that future versions of Billie will not encroach on design advisory work as readily as they did on basic customer service.

No comments yet. Be the first to share your thoughts!