Darren Acemoglu, Simon Johnson and James A. Robinson received this year’s Nobel Prize in Economic Sciences for their contributions to proving the importance of public institutions to a country’s prosperity.
KEY FACTS
The prestigious prize, officially known as the Sveriges Riksbank Prize for Economic Sciences in Memory of Alfred Nobel, is the last prize awarded this year and is worth SEK 11 million ($1.1 million).
This year’s laureates showed that one of the explanations for differences in countries’ prosperity is the social institutions introduced during European colonization. Inclusive institutions were often introduced in countries that were poor at the time of colonization, which over time led to general prosperity for the population. This is an important reason why former colonies that were once rich are now poor and vice versa.
Introducing inclusive institutions would create long-term benefits for everyone, but extractive institutions provide short-term gains for those in power. As long as the political system ensures they retain their control, no one will trust their promises of future economic reforms. According to the laureates, this is the reason why there is no improvement.
“Reducing the huge income gaps between countries is one of the greatest challenges of our time. The laureates have demonstrated the importance of public institutions in achieving this,” said Jakob Svensson, Chairman of the Economic Sciences Prize Committee.
“Societies with poor rule of law and institutions that exploit the population do not generate growth or change for the better,” the prize’s organizers add on their website.
TANGENT
Darren Acemoglu and Simon Johnson work at MIT, while James Robinson is at the University of Chicago.
Acemoglu and Johnson recently collaborated on a book researching technology through the ages that demonstrates how some technological advances are better at creating jobs and spreading wealth than others.
KEY STORY
The Economics Prize is not one of the original science, literature and peace prizes created by the will of dynamite inventor and businessman Alfred Nobel and first awarded in 1901, but is a later additional prize established and funded by the Central Bank of Sweden in 1968.
Past recipients of the award include a number of influential thinkers such as Milton Friedman, and John Nash – played by actor Russell Crowe in the 2001 film A Beautiful Mind, and former US Federal Reserve Chairman Ben Bernanke.
Last year, Harvard economic historian Claudia Goldin won a prize for her work highlighting the causes of pay and labor market inequality between men and women.
When AI Agents Start Shopping For Your Clothes: Fashion’s Agentic Commerce Challenge
Agentic AI can book your flight and reorder your coffee. But fashion shopping runs on browsing, inspiration, and bodies that don’t come in standard sizes. That combination is proving far harder for autonomous agents to crack.
The Promise Meets Its Hardest Category
Late last year, we covered how agentic commerce is reshaping global transactions. The illustration was crisp: tell an AI to find the cheapest red-eye flight from Singapore to Tokyo under $500, and it searches, compares, books, and pays. Done. The entire purchase happens inside a single conversation.
Flights are standardized products. A seat is a seat. A price is a price. The agent’s job is clear, the criteria measurable, the outcome binary. But what happens when the AI agent needs to buy you a dress for a wedding in Mykonos?
Fashion is where agentic commerce runs into a wall. And the reasons go deeper than most industry commentary acknowledges.
Fashion Is A Browsing Category, Not A Searching Category
When someone shops for electronics, they typically know the product. “Samsung Galaxy S26, 256GB, best price.” The intent is specific, the comparison is numerical, and an AI agent can handle it without breaking a sweat.
Fashion works differently. Most consumers don’t know what they want when they start shopping for clothes. They browse. They scroll. They stumble onto a jacket they didn’t know existed and suddenly rethink the entire outfit. This isn’t a flaw in how people shop. It’s the point.
Academic research confirms what anyone who has ever spent 40 minutes on a fashion app already knows: online clothing shopping is dominated by what researchers call “diversive exploration” — browsing for enjoyment and discovery, distinct from goal-directed search. The behavior is hedonic, not utilitarian. People don’t just want the product. They want the process.
The numbers back this up. According to McKinsey’s State of Fashion 2026 report, shopping-related searches on generative AI platforms grew 4,700% between 2024 and 2025, with AI supporting “inspiration and product comparison” — especially in fashion, where choice abounds. Consumers are using AI to discover, not to delegate. A separate Bain & Company study from April 2026 found that 44% of US online buyers now start their journey in an LLM or split between AI and traditional search. But in fashion specifically, 46% use AI for “discovering new products and getting inspired,” while usage drops sharply as activities move closer to checkout and payment.
An AI agent can book a flight autonomously because the consumer’s intent is clear. In fashion, the intent is often vague, “something for summer”, or absent: “I’m just looking.” You can’t delegate browsing to an agent. Browsing is the experience.
Even When You Know Exactly What You Want
Suppose a consumer does have a specific goal. They want a pair of Camper Pelotas in size 42. Straightforward enough for an AI agent, right?
Not quite.
A size 42 in Camper is not a size 42 in Nike, which is not a size 42 in Adidas. There is no universal sizing standard in fashion. Every brand calibrates differently, and some brands are inconsistent across their own product lines. An AI agent that confidently orders the “right” size has roughly a coin-flip chance of getting it wrong in certain categories. European fashion return rates hover between 25% and 40%, with size and fit issues accounting for more than half of all returns, according to Statista and European e-commerce industry data. In Germany, the practice of “bracketing”, ordering three sizes of the same item to try at home, pushes online fashion return rates above 44%.
Then there’s the visual dimension. A flat product photo in an AI chat window doesn’t replicate what happens when a consumer sees a shoe alongside ten alternatives on a comparison grid. Context matters. Styling matters. The way a sandal looks next to a linen dress matters. Pinterest’s visual search technology has driven a 387% revenue increase for participating merchants, and visual search users convert at rates 73% higher than text-based searchers, according to industry data tracked by eCommerce Times. Platforms like Spangle are proving that AI-powered visual personalization lifts revenue per visit by up to 50%.
There’s a final paradox. Price comparison absolutely works in fashion — the same branded shoe can differ by 30% across retailers. But consumers also compare across products. “Do I want the Camper or the Clarks?” That requires visual side-by-side browsing, and current AI agents can’t replicate it well. They’re designed to return a result, not to facilitate a process.
The Infrastructure Gap
For AI agents to operate autonomously in fashion, they need structured, real-time data: normalized product attributes, cross-retailer pricing, size mapping, availability signals, and brand reliability scores. This infrastructure barely exists.
Consider how hard this is even in simpler categories. Cyprus’s government-backed e-Kalathi grocery comparison platform launched with the goal of transparent supermarket price tracking. Within months, the Cyprus Consumer Association flagged accuracy problems — pricing inconsistencies, incomplete product coverage, misleading comparisons. And that’s groceries, where a bottle of milk is a bottle of milk.
Fashion is orders of magnitude harder. Product feeds arrive in dozens of incompatible formats. A “navy blue slim-fit cotton shirt” from one retailer might be listed as a “dark blue fitted cotton top” from another — same product, entirely different data. Normalizing that across thousands of products from dozens of retailers requires purpose-built AI pipelines. Stylino, a Cyprus-based fashion price comparison engine, processes feeds from 65+ retailers and uses AI to match and deduplicate over 385,000 products into a single searchable catalogue. Building that kind of data layer took months of custom engineering — and it’s the sort of plumbing that agentic commerce will eventually need to function in fashion.
On the visual side, companies like Aiuta are using AI to generate styled product imagery and virtual try-on experiences, addressing the content bottleneck that currently limits how well any automated system can present fashion to consumers. These building blocks, structured data, visual content, size intelligence, will eventually form the infrastructure layer that agents plug into. But we’re early.
The Likely Sequence
Fashion won’t leap from browsing to fully autonomous purchasing. The transition will happen in stages, and each stage suits a different kind of AI intervention.
First, consumers browse and discover. This is visual, emotional, and social. It won’t be delegated to an agent anytime soon, because delegation defeats the purpose. Second, AI helps compare prices and availability across retailers — this is already happening and provides genuine value. Third, AI monitors price drops, tracks wish lists, and sends alerts when a saved item goes on sale. Useful, but still decision-support rather than decision-making. Fourth, AI executes purchases on known, pre-approved items: reorders, basics, and items the consumer has bought before in the right size.
Only that last step is truly “agentic.” And it applies primarily to commodity fashion: underwear, socks, a replacement white t-shirt, not to the discovery-driven shopping that accounts for most fashion spending. McKinsey’s European agentic commerce research confirms this sequencing: AI is being adopted first as a “decision-support layer, compressing research, comparison, and synthesis,” with usage declining as activities move closer to execution.
Here’s the uncomfortable truth for the agentic commerce narrative: the fitting room is where most fashion decisions actually happen. It’s physical. It’s emotional. Sometimes it involves a friend outside the curtain saying, “absolutely not.” AI agents are exceptional at finding you the cheapest red-eye to Tokyo. They are not standing in that fitting room mirror with you. The agent who wins in fashion won’t be the one who buys for you. It’ll be the one that helps you see better: more options, better prices, smarter comparisons, while you keep making the call.
Manage Cookie Consent
We use cookies to enhance your browsing experience. By clicking "Accept All", you consent to our use of cookies. Privacy Policy
Functional
Always active
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
Preferences
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
Statistics
The technical storage or access that is used exclusively for statistical purposes.The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
Marketing
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.
