For most of my life, a visit to IKEA meant being on your own. You followed the arrows on the floor, wrote numbers down with a tiny pencil, and carried your decisions to the warehouse yourself. That was the deal: low prices in exchange for doing the work. Nobody walked you through your living room. If you wanted advice, you hired an architect, and an architect was a different price category, a different world.
Then something changed, quietly, and I think it's one of the most interesting business stories of the decade.
An IKEA co-worker at the Planning Studio in Melbourne. Courtesy of Ingka Group.
A machine that answers, a person who advises
In 2021, Ingka Group, the largest IKEA retailer, introduced an AI assistant called Billie to its customer service. Billie took over the routine questions: where is my order, how do I return this, is this in stock. Over two years, it handled 47% of customer queries. In most companies, a sentence like that ends with the word layoffs.
At IKEA, it ended differently. Instead of letting go of the people whose questions Billie now answered, the company trained 8,500 call center workers to become interior design advisers. They learned to plan rooms, to listen to how people actually live, and to suggest the pieces that would make a small apartment work. Customers can now book a video call and receive a floor plan, 3D visuals, and a shopping list. In the 2022 financial year, remote interior design brought in 1.3 billion euros, about 3.3% of total sales. The target is 10% by 2028.
Asked by Reuters whether AI would reduce headcount, Ingka's global people and culture manager answered: "That's not what we're seeing right now." I don't read that as a promise. I read it as a choice someone made, and keeps making.
It's not the same as hiring an architect. An architect is the premium option, and always will be. But IKEA was never meant to be premium. It was built for people, for first apartments and tight budgets, and now those people get something that used to belong only to clients who could pay for it: a professional eye. The machine took the questions. The people got the conversations.
The loom that wove flowers
Fear of the machine is older than electricity, and fashion has a special place in its history.
In 1804, a weaver's son from Lyon named Joseph-Marie Jacquard built an attachment for the loom that read patterns from a chain of punched cards. Each card told the threads what to do. Brocades and damasks that once needed a master weaver and a helper pulling cords by hand could now be woven by one person, row by row, from a program.
The silk weavers of Lyon understood immediately what this meant for them. They burned looms that had been put into production and attacked Jacquard himself. By 1812, about 11,000 of his looms were working in France anyway.
Portrait of Joseph-Marie Jacquard, woven in silk on a Jacquard loom, Lyon, c. 1839. Public domain, via Wikimedia Commons.
Twenty years later, the city's silk workers, the canuts of the Croix-Rousse hill, rose up twice, in 1831 and 1834. Their grievance was pay more than any single machine: the price of their work kept falling, while the silk they wove dressed the richest people in Europe. They marched under a sentence I can't forget: live free working, or die fighting.
Here is the part I love most. Jacquard's punched cards did not stay in the silk trade. In London, Charles Babbage borrowed them for his Analytical Engine, the first design for a general-purpose computer. In 1843, Ada Lovelace, who wrote the first program for that machine, explained it like this: "We may say most aptly that the Analytical Engine weaves algebraical patterns just as the Jacquard loom weaves flowers and leaves."
Ada Lovelace, portrait by Alfred Edward Chalon, c. 1840. Science Museum, London. Public domain.
So the computer has a grandmother in a silk workshop, and one of the first fears of automation was a weaver's fear. The thread from the loom to the algorithm is a straight one. Last week, in The Glass Between Us, I wrote about the Luddites, who broke machines in England between 1811 and 1816. They were not against progress. They were afraid of becoming unnecessary. Every industrial revolution has carried that fear, and every one was also, in a way, a revolution of people nobody asked.
Keynes's grandchildren
In 1930, in the middle of the Depression, the economist John Maynard Keynes wrote a short essay called "Economic Possibilities for our Grandchildren." In it, he gave a name to something people already felt: technological unemployment, the loss of work faster than we can find new uses for labor. He called it a temporary phase of maladjustment. His grandchildren, he predicted, would work around fifteen hours a week and face a very different problem: how to use their freedom well.
We are those grandchildren, more or less. Nobody I know works fifteen hours a week. The machines arrived. The freedom, for most people, didn't. What happened to the time technology saved depended, every single time, on what someone decided to do with it.
That is the question I keep returning to. Not whether AI will replace tasks. It will, and in many places it already has. The question is who decides what happens to the people whose tasks are gone, and what they decide.
Modern Times (1936), written and directed by Charlie Chaplin, cinematography by Roland Totheroh and Ira Morgan. © Roy Export S.A.S.
The year of the efficient cut
The last two years have given us a lot of answers, and most of them were the same answer.
In 2025, companies in the United States directly cited AI as the reason for about 55,000 job cuts, according to the outplacement firm Challenger, Gray & Christmas. That is more than twelve times the number from two years earlier. This year, the announcements kept coming. Block said it would shrink from about 10,000 people to 6,000. Pinterest cut 15% of its staff to support what it called an AI-forward strategy. PayPal pointed to 1.5 billion dollars in savings from AI and automation. Not every one of these cuts is really about AI. Some economists suspect that AI has simply become a more flattering word for an ordinary layoff. But the story a company chooses to tell says a lot about it.
The most instructive story belongs to Klarna. In early 2024, the Swedish payments company announced that its AI assistant was doing the work of 700 customer service agents. A year later, its chief executive admitted that the strategy had produced lower quality, and the company began hiring people again. "Really, investing in the quality of human support is the way of the future for us," he told Bloomberg. It took a year and a lot of unhappy customers to learn something IKEA had planned from the start.
Peace is also a business decision
I am not writing this as an economist or a sociologist. I am writing as a human being who watches what happens around her, and feels it.
When a company replaces people with AI, its job does not end at the spreadsheet. It also has to think about what happens to those people. I don't see that as charity, just common sense. Thousands of people who suddenly feel unnecessary don't disappear. They go home, to their families and their neighborhoods, carrying anger and fear, and that mood finds its way back into the streets, into elections, and into the way people look at brands. Last week, I wrote about Paris and the shop windows smashed during the riots. Mass layoffs won't bring peace. A person who has lost their work and their place in the world has very little left to lose.
A human being wants to live with dignity. Work is not only income. It is the feeling that you are needed, that your hands and your eyes still matter to someone. Take that away from enough people at once, and no quarterly result will balance it.
There is another way, and it is not naive. You let the machine do what it does best: the repetitive, the predictable, the questions with one right answer. And you move people toward the places where another human being is truly needed. IKEA's numbers show that this is not a trade-off between kindness and profit. The two grew together.
I think about this personally, too. Anyone who has built their own brand knows that in the beginning, you do every job yourself. There is no team yet, only you and the work. It is exhausting, and it teaches you something priceless. When the resources finally grow and those jobs go to other people, you know each of them from the inside. You know what a good result looks like, how long it really takes, and where it usually goes wrong. You can write a brief that someone can actually work from, and you can hold the result to it. AI will change many of these jobs. The person who has done them will still be the one who knows how to lead them.
What only a human eye can do
So where are the places that still need a human?
Advice, first of all. Not information, which a machine gives you in a second, but advice: the moment when someone looks at you, at your life, your budget, and your taste, and says, this, not that. Then taste. A machine can recommend what is popular. A person can tell you what is right for you, which is sometimes the opposite. And trust. We forgive a person who makes a mistake and explains it. We rarely forgive a system.
Then there is care: the alteration done properly, the repair, the call after a purchase, the handwritten card. And craft, the work of hands, which in fashion has always been the real luxury. In the Umbrian village of Solomeo, Brunello Cucinelli built his company around what he calls humanistic capitalism, and opened a school where young people learn crafts such as tailoring and mending. You can find that idealistic. You can also notice that it is one of the most consistent brands in luxury.
The vineyard in Solomeo, Umbria. Courtesy of Brunello Cucinelli.
The sales floor in luxury
Which brings me to fashion, and to luxury in particular, where this question is about to become urgent.
Luxury has had difficult years. Prices rose much faster than people's patience. According to Bain and Altagamma, the luxury market lost around 50 million customers between 2022 and 2024. Add the social mood, the growing distance between those who buy and those who watch through the glass. Add AI, which can already write product descriptions, answer questions about delivery, and suggest the bag that matches your shoes.
The temptation is obvious: automate whatever can be automated, cut costs, protect the margin. Some of that will make sense. Nobody needs a person to tell them where their parcel is.
But when a luxury brand takes people out of the moment of choice, it takes away the very thing it sells. Someone who pays thousands for a coat wants more than fabric. She wants to feel that a person with a good eye actually saw her. That is where I would invest the time AI frees: in training sales associates not only to sell, but to advise like a stylist and close like a good salesperson. Someone who knows the collection, the cut, the fabric, and the client's wardrobe, and can say, honestly: this is not for you, but this is. In my experience, that sentence builds more loyalty than any discount.
Pretty Woman (1990), directed by Garry Marshall, cinematography by Charles Minsky. Julia Roberts as Vivian Ward. © Touchstone Pictures.
Everyone remembers the scene where Vivian walks back into that boutique, arms full of shopping bags, and says: big mistake. Huge. The saleswomen didn’t lose a sale because of the price. They lost it because they didn’t see her.
For years, these two roles were treated as separate. The stylist was creative, the salesperson was commercial, and the two rarely met. IKEA's results suggest that they belong together. Advice sells. A person who feels well advised comes back.
What the houses are already doing
The good news is that some of the biggest names in luxury seem to be getting there too, each in their own way.
At VivaTech in Paris in June 2025, Louis Vuitton's vice president of global retail excellence described how the house uses AI for its client advisors: to help them decide which clients to contact, when, and with what, and to take over the administrative work that eats their day. The goal, she said, is to "give them more time to really focus on high added value interactions." One pilot uses generative AI to draft personal thank-you notes in a tone matched to each client. "It's really about simplifying the life of our client advisors." Notice the direction. The machine works behind the advisor, not instead of her.
Across LVMH, an internal assistant called MaIA now answers more than two million requests a month from around 40,000 employees, mostly translations, summaries, and drafts. In the stores of several maisons, AI tools bring up a client's history and preferences, so that an advisor can prepare a conversation instead of starting from zero. The same group has run, since 2014, the Institut des Métiers d'Excellence, a free work-study school that has trained more than 3,300 apprentices in leather goods, couture, jewelry, and watchmaking, and also in selective retail, the art of selling. In 2025 alone, more than 4,500 jobs were offered in these trades in France, and a whole generation of artisans is expected to retire by 2030. I like that the same group invests in AI and in apprentices at the same time. Somebody there understands that one doesn't cancel the other.
Online, the picture is more complicated, and more interesting. In September 2025, Ralph Lauren launched Ask Ralph, an AI stylist built with Microsoft and designed to feel like talking to a stylist in a store. You describe the occasion, and it assembles a look from the collection. For a customer at midnight, in a small town far from any flagship, that is a gift. But it also raises the question this whole Note is about. Is the digital stylist a bridge to the human one, or a replacement? The answer will not come from the technology. It will come from what the company decides to do with its people.
Stitch Fix, the American styling service, has been answering that question for years. Its algorithms narrow the choice using data, and human stylists make the final call. In 2025, it added an AI style assistant and, at the same time, made it easier for clients to message their stylists directly. Its chief executive has said that AI can supplement stylists, but not replace them. A company built on algorithms decided that the human was the part worth keeping.
And then there is the simplest argument of all. At Mytheresa, one of the most successful luxury e-commerce platforms, top customers made up 39% of the business in its 2023 financial year, and their spending grew by more than 30%. Those clients are not won by a chatbot. They are won by personal shoppers, private events, and a trip to Portofino to launch a capsule collection. In luxury, the best clients are relationships, and relationships are made by people.
Fewer stores, better people
The news this fall makes all of this more urgent. According to FashionUnited, Kering has closed more than a hundred stores this year and plans to shut 126 boutiques by the end of it. Luxury is not disappearing, but it is concentrating: fewer doors, each one carrying more of the brand's weight. When there are fewer stores, every person in them matters more.
A boutique with fewer, better trained, and better paid advisors, supported by tools that take care of the paperwork, is not a sentimental idea. It may be the only model that holds. The client who walks in today has already seen the product online, compared prices, and maybe checked what it would resell for. The only thing she can't get on her phone is a person who knows her, remembers her last purchase, and has the courage to say: not this one, wait for the next delivery, it was made for you.
In practice
If you run a brand, a store, or a team, and AI is already on the table, here is how I would think about it.
I would start by asking what the machine should take. Not everything that can be automated should be, but the repetitive and predictable work, the order status, the size chart, the return label, can go to a system without anyone missing it.
Then I would ask who is doing those tasks today. These people know your products, your customers, and your problems better than any consultant. Before anyone writes their names on a list, it is worth asking what they could do with the time the machine gives back.
And finally, where a human eye actually makes a difference. In fashion, the answer is usually advice, styling, fitting, alterations, after-sales care, and the long relationship with a client. Train people for that. Measure them by whether clients come back, not only by what they sold today.
It may sound soft. It isn't. IKEA has the numbers to show it.
The human who stays
Lovelace said that the machine weaves algebraic patterns the way the loom weaves flowers and leaves. Almost two centuries later, the machine weaves nearly anything. It can describe the flower, photograph it, and generate a thousand versions of it. What it still can't do is stand next to you and know which one belongs on your table.
I believe we meet each other for a reason, even at a store counter, even on a video call about a kitchen. A person in front of you changes the energy in the room. That is not something you can automate, and I don't think the companies that try will be the ones people remember.
The future I would like to see is not one where machines replaced people. It is one where machines took the questions, and people kept the conversations.
xx, Laura
