A viral LinkedIn post recently presented IKEA as a company that replaced 8,500 customer service roles with artificial intelligence and became dramatically richer because of it.
The real story is more useful than the viral version.
IKEA did not report laying off 8,500 people because of a chatbot. Ingka Group, the largest IKEA retailer, says its artificial intelligence assistant Billie resolved about 47 percent of the inquiries it received from 2021 through 2023. During that period, the company retrained 8,500 remote customer support coworkers for work that included remote interior design, digital sales, relationship building and more complex customer questions.
IKEA automated routine work, but it did not treat the people performing that work as disposable.
That is the hopeful part of the story.
It is also the part we should not assume every employer will repeat.
At a glance
What IKEA actually reported
From 2021 through 2023, Billie handled 3.2 million customer interactions and resolved about 47 percent of the inquiries sent to it. Ingka says that work produced nearly 13 million euros in savings.
The company also reported 1.3 billion euros in sales through its remote customer meeting channel during its 2022 fiscal year.
Important context: The 47 percent figure refers to inquiries Billie received, not every customer interaction across IKEA. The 1.3 billion euros represents remote channel sales, not profit and not revenue proven to have been created by artificial intelligence.
The clearest lesson is simpler: routine questions moved to software, while thousands of employees were trained for work that required more from them.
That is transformation rather than mass dismissal.
The future will not always arrive that gently
It is tempting to look at IKEA and decide that artificial intelligence will simply remove repetitive tasks while everyone moves into more interesting work.
That outcome is possible.
It is not automatic.
Retraining 8,500 people takes money, time, managers, clear new roles and a company willing to absorb transition costs. Smaller employers may not have those resources. Other companies may have them and still choose a faster route: eliminate tasks, slow hiring and ask the remaining staff to carry more work.
Fewer openings
Displacement can begin when an entry level role is never posted after routine work is automated.
Less work
Freelancers and hourly employees may see fewer assignments or fewer scheduled hours before a company announces layoffs.
Higher barriers
A familiar role may return with new technical requirements that existing workers were never given time or support to learn.
The unemployment rate can remain calm while individual workers feel the ground moving beneath them.
What the evidence says right now
The International Labour Organization estimates that one in four workers worldwide is in an occupation with some exposure to generative artificial intelligence. It also says transformation is more likely than complete replacement.
That is an important guardrail against panic. Exposure is not the same as a forecast that a job will vanish.
At the same time, the risks are not imaginary. Recent evidence points to pressure on entry level hiring, freelance work and office support tasks even while broad labor market displacement remains limited.
The concern is not that every exposed worker loses a job tomorrow.
The concern is that opportunity may narrow first for the people with the least experience, the least bargaining power and the least access to paid training.
Why this matters in Southern Utah
Office and administrative support is the largest occupational group in the St. George metropolitan area, representing 13.1 percent of local employment in May 2025, according to the United States Bureau of Labor Statistics. The mean local wage for that group was 22 dollars and 40 cents per hour.
Many of those jobs include scheduling, documentation, routine customer communication, data entry and information retrieval. Those are exactly the kinds of tasks artificial intelligence can increasingly accelerate or absorb.
That 38 percent figure is an exposure classification, not a forecast. It does not mean 38 percent of Utah jobs will disappear. It means a significant share of workers may see duties, hiring prospects or skills requirements change.
A transition of that size deserves planning before the disruption reaches a family paycheck.
A worker plan should be part of every automation plan
Businesses have every reason to use tools that improve service and reduce repetitive work. Refusing useful technology is not a serious strategy.
But measuring success only through savings leaves out the people who made the old system work.
Questions every employer should answer
IKEA had answers for many of those questions. That is why its story deserves attention.
Our take
Artificial intelligence is not inherently the enemy of workers.
Indifference is.
The IKEA example shows that technology can remove routine work without automatically removing the people who performed it. It also shows how much deliberate effort that outcome requires.
We should celebrate employers that retrain workers rather than discarding them. We should also be honest that voluntary corporate goodwill is not a complete plan for a labor market changing at this speed.
The future of work should not be judged only by how many questions a chatbot answers, how much money a company saves or how quickly a task disappears.
It should also be judged by where the people doing that work are able to land.
Progress is real only when workers have a way forward.
Source note
This commentary responds to Chris Donnelly’s LinkedIn post.
The IKEA figures and workforce details come from Ingka Group’s account of Billie and remote selling. The featured photograph is an official Ingka Group press image from that report.
Workforce exposure context comes from the International Labour Organization global index and its 2026 evidence review.
Local employment figures come from the United States Bureau of Labor Statistics report for St. George. Utah context comes from the Utah Department of Workforce Services.
