The Swedification of Everything

It's hard not to like Swedish products, and similarly hard not to like some of their most prominent business philosophies. Here I will draw a parallel with what these very particular businesses have in common (and how they differ) from the current effects we are assing with AI and day to day work.

A few years ago I was invited to speak at an event hosted inside the H&M headquarters in Stockholm. Somewhere between the coffee and the presentation, a senior executive turned to me with the easy confidence of someone who already knew the answer, and asked: is H&M your go-to store? It was a disarming question, and I have been thinking about it ever since. The goal, I realised later, was not to sell me a specific jacket or shirt. The goal was to become the place I go first, before considering anywhere else — to win the default position in my wardrobe, and therefore in my monthly spending. That conversation, more than any consulting deck I have read about platform economics, clarified for me what a certain kind of company is actually trying to do. And it clarified why the AI tools we are now adopting at scale feel, on closer inspection, so structurally familiar.

Sweden has exported to the world a particular model of consumer capitalism that is worth examining carefully, because it is arriving again, in a new form, through artificial intelligence. The model is elegant. It takes things that used to be accessible only to a cultivated elite — designed furniture, well-curated fashion, an edited music library, a pleasant place to sit and work — and makes them available to almost everyone, at a price most people can absorb, with a standard that hovers comfortably above average without ever threatening to reach the top. The experience is pleasant, the aesthetics are clean, the friction is low. And over time, the relationship compounds: a loyalty card here, a subscription there, a slow expansion of the share of your attention and your wallet that belongs to one brand, then two, then a small constellation of them.

IKEA is the template. It democratised the Bauhaus dream of accessible modern design and, in doing so, made something genuinely new: a furniture experience in which the buyer is also the assembler, the curator, and — through the act of putting the pieces together in their own home — a kind of co-author. There is a small psychological investment in an IKEA bookshelf that a fully assembled alternative would not produce, and that investment is part of what makes the product work. The materials, of course, are often thinner than one would like. The pieces may last five years instead of fifty. But the deal is honest on its own terms: accessible design, built with you, for now. The same logic animates H&M, which has brought the visual grammar of fashion history to high streets across the world at prices that would have seemed impossible a generation ago, at the well-known cost of a fast-fashion supply chain whose environmental and labour records are notoriously difficult to defend. Spotify continued the pattern into culture itself, democratising access to nearly all recorded music ever made while, at the same time, restructuring the economics of being a musician to the point where most artists now earn meaningfully less from a stream than they did from a CD, and where the platform has begun to produce the lifestyle wrapper — the yearly retrospective, the algorithmic playlist, the mood-based session — within which listening now happens.

When you travel north from Helsinki toward Stockholm, as I have done more than once, another Swedish icon becomes impossible to ignore. Espresso House has colonised the Nordic urban centre with a particular kind of domestic calm: low lighting, soft plants, a visual vocabulary borrowed from the Danish notion of hygge, a communal anonymity in which you do not need to know anyone at the next table but you feel, somehow, among people of your kind. The coffee is genuinely good. The pastries are a little bulky, a little uneven, not quite at the level a serious café in Lisbon or Vienna would tolerate. But the proposition is not perfection; the proposition is that you can work here, or meet a friend here, or sit alone here, and the experience will be consistently pleasant in a way that almost no independent alternative can match at the same price. A loyalty card hovers in the background, quietly increasing over the year.

Swedes have a word for this disposition, and the word matters. Lagom is usually translated as "not too much, not too little," or "just the right amount," and it is often marketed abroad as a lifestyle aesthetic — minimalism with better lighting. But lagom is more interesting than the lifestyle packaging suggests. Its etymological roots trace to the idea of what is appropriate for the group, what is fair when a shared resource is passed around, what allows a collective to function without anyone taking too much or being left with too little. It is a structural disposition before it is a personal one. And when translated into business design, lagom produces exactly the pattern IKEA, H&M, Spotify, and Espresso House have each exported: a product that is good enough for most people, affordable enough for most people, accessible enough for most people, and that resists the twin temptations of luxury exclusivity on one side and bargain-basement disposability on the other.

The reason this model is worth attending to now is that AI tools are built on the same structural logic, at a scale the Swedish exporters never quite managed. A subscription to a large language model costs roughly what a Spotify family plan costs, and delivers a quality of drafting, summarising, and structuring that would have been available a decade ago only to people who could afford a research assistant, a junior copywriter, or a branding consultant. The floor of professional output has risen — measurably, across most of the economy — and the cost of crossing that floor has fallen to almost nothing. This is, in its own way, a democratisation of cultural capital, of the kind Pierre Bourdieu described as the most durable form of social inequality: the internalised dispositions, the aesthetic confidence, the fluency in institutional registers that separate those who move easily through professional life from those who must work twice as hard to do the same. AI, at least on the surface, distributes a version of that fluency to anyone willing to pay twenty euros a month.

What the tools produce, however, is recognisably uniform. The websites built with Claude are good websites. They are clean, legible, and well-structured, and they solve problems that used to require a design team. But they all feel, to a trained eye, a little bit the same — as if they had emerged from a shared gravitational centre that the model cannot help but orbit. Writing assisted by GPT becomes clearer, more balanced, better edited, and systematically free of typos, which is not nothing. It also tends toward a particular middle register, a cadence of competence that is consistently above average but rarely far above it. Byung-Chul Han, in Saving Beauty (Han 2015), described the aesthetic signature of our moment as smoothness: surfaces without resistance, outputs without rupture, experiences engineered to elicit the small affirmation rather than the unsettling insight. Generative AI is, in this sense, the industrial realisation of the smooth. It produces the statistical mean of its training corpus, and the statistical mean is almost always pleasing, almost always reasonable, and almost never strange.

Gilles Lipovetsky's framework of artistic capitalism is useful here, because it describes the longer arc that AI is accelerating rather than inaugurating. In L'esthétisation du monde (Lipovetsky and Serroy 2013), he argued that late consumer capitalism has saturated every domain of life with aesthetic experience — that the market has made beauty, design, and curated sensation available at a scale no previous civilisation could have imagined. The Swedish exporters are masterful players of this game: they deliver aesthetic pleasure at industrial volumes, at a quality calibrated to feel meaningfully above commodity without ever asking the consumer to develop a connoisseur's palate. Lipovetsky's contribution is to note that this saturation does not abolish distinction; it displaces it. When everyone has access to reasonable design, reasonable fashion, reasonable music discovery, the grounds of social differentiation migrate elsewhere — to time, to attention, to the handful of experiences that cannot yet be commodified. AI moves that frontier forward again, by commodifying the production of competent professional work itself.

The structural novelty of what is happening now, and the reason the Swedish analogy only takes us part of the way, is that AI personalisation operates at a scale and a granularity that no prior consumer platform has matched. Spotify personalised playlists to my listening history; IKEA personalised its catalogue to my living room only in the loosest sense. A language model, by contrast, personalises each sentence to the specific context I have just typed, and does so for hundreds of millions of people simultaneously, each receiving an output shaped to feel as if it had been written for them. This is mass personalisation in a sense the word has not previously meant. It is the industrial production of the artisanal feeling — the sense that something has been made with you in mind — at a scale that no human workshop could ever approach. And it rests, structurally, on the same bargain the Swedish exporters have long offered: you get something that is good enough, tailored enough, pleasant enough, in exchange for a modest recurring payment and a quiet expansion of the platform's share of your everyday life.

The trade-offs are worth naming directly, because they are not the trade-offs critics usually name. The familiar critique says that AI will replace jobs, hollow out creative work, or degrade public discourse. Those concerns are real, but they miss what the Swedish exporters have already demonstrated about how this kind of platform actually reshapes economies. The deeper trade-off is about the middle. When the floor of output rises, the middle thickens. More people can produce competent work, which means competent work ceases to be a meaningful differentiator. The BCG research on Nordic AI adoption published in early 2026 makes this point in a different register: companies that have poured investment into off-the-shelf AI productivity tools are seeing gains of ten to twenty percent in targeted activities, but they are not pulling away from competitors who are using the same tools (Boston Consulting Group 2026). The tools are too evenly distributed to produce durable advantage. What they produce is a new industrial baseline, and like all industrial baselines, it is a ceiling for many and a floor for few.

The second trade-off is about accumulated dependency, and it is the one the H&M executive's question pointed toward. Is H&M your go-to store? is the question every platform is now, in some form, asking. Is Spotify your default for music? Is IKEA your default for furnishing a new apartment? Is Claude, or ChatGPT, your default for thinking through a problem? Each default, once established, produces a compounding relationship. The loyalty card matures. The playlist history becomes too costly to rebuild elsewhere. The chat history becomes a substrate of your professional practice. The share of wallet quietly grows, not because the platform is extortionate but because it has become genuinely useful, and genuinely useful things are genuinely expensive to leave. The token bill rising each month is structurally identical to the slowly accumulating IKEA furniture that would now take a full day to disassemble if you moved. Neither platform needs to lock you in. The accumulation does the work.

The third trade-off, and the one that deserves the closest attention from designers and strategists, is the narrowing of the expressive range at the population level. Individual creativity has likely been enhanced by these tools: a person who used to struggle with a blank page now produces reasonable drafts, and the research on AI-assisted writing shows genuine gains in individual output quality (Noy and Zhang 2023). But the collective space of what gets produced narrows, because everyone is drawing from roughly the same distributional centre of the model's training data. Doshi and Hauser (2024), in a study published in Science Advances, documented this pattern clearly in creative writing: participants using generative tools produced individually stronger stories, but the stories became measurably more similar to one another. The aggregate creativity of the group declined even as the individual average rose. This is the Spotify effect scaled to every domain of symbolic work. A richer average, a narrower distribution, a thinner edge.

What the Swedish model has always handled gracefully is the question of what lagom is for. The point of the disposition is not mediocrity; it is sufficiency calibrated to the collective. A country that has internalised the idea that enough is a real quantity — not a euphemism for less, not a failure of ambition — can build institutions and products that serve most people decently without needing to promise anyone everything. The danger in the current AI moment is that the Swedish pattern is being exported without the Swedish intellectual infrastructure around it. We are receiving the sufficient tools without the cultural commitment to ask what they should be sufficient for. A product can be lagom in the Swedish sense only if the society in which it circulates has agreed on what a good-enough outcome looks like. Absent that agreement, lagom collapses into its commercial shadow: a standardisation that serves whoever sets the defaults, rather than the collective that was meant to benefit.

This is the point at which the designer, the strategist, and the product professional acquire a clearer kind of responsibility. The tools will continue to improve, and the share of everyday professional work that an off-the-shelf model can handle competently will continue to grow. That is not, on its own, a problem — it is genuinely useful, and the floor it raises is a floor many people needed raised. The harder question is where the edge now lives, and who is building it. The Swedish exporters succeeded because, beneath the pleasant surface, someone had made demanding decisions about what to leave out, what to refuse, and what to stand behind. IKEA's catalogue is edited. H&M's collections are curated, even when the quality is not. Spotify's original product was a decisive bet on streaming economics at a moment when the industry disagreed. The platforms that have endured are the ones whose lagom was the outcome of taste, discipline, and a specific view of what the collective needed. The platforms that will endure in the AI era will be built the same way — and so will the practices of the people using them. A drafted document, a generated website, a synthesised strategy is not yet a decision. It is a starting point from which the two irreducibly human questions still need to be answered: what needs to happen here, and are we willing to stand behind it. Everything between those two questions has been Swedified. The questions themselves have not.

References

Boston Consulting Group (2026). The Nordic AI Inflection Point: Value Creation or Value Bubble? BCG.

Bourdieu, P. (1979). La Distinction: Critique sociale du jugement. Paris: Éditions de Minuit.

Doshi, A. R. and Hauser, O. P. (2024). "Generative AI enhances individual creativity but reduces the collective diversity of novel content." Science Advances, 10(28).

Han, B.-C. (2015). Die Errettung des Schönen. Frankfurt am Main: S. Fischer Verlag. [English translation: Saving Beauty, Polity Press, 2018.]

Lipovetsky, G. and Serroy, J. (2013). L'esthétisation du monde: Vivre à l'âge du capitalisme artiste. Paris: Gallimard.

Noy, S. and Zhang, W. (2023). "Experimental evidence on the productivity effects of generative artificial intelligence." Science, 381(6654), 187–192.