I read a lot of industry reports. It is part of the job, and honestly I enjoy it. But over the past few months I keep having the same slightly funny reaction. Report after report sets out to describe where fashion commerce is heading, and what they end up describing is the thing we already built.
The latest one to do it is McKinsey.
Their new article on Europe’s e-commerce agenda puts a big number on the table: online revenue across the five largest European markets, Germany, the UK, Spain, Italy and France, could reach 600 billion euros by 2029, growing around 6 percent a year. The number is nice, but it is not the interesting part. The interesting part is the engine. Roughly 38 percent of European shoppers already use generative AI to research products and decide what to buy. They are handing real tasks to AI now. Compare these prices. Reorder that. Build me a cart that fits my budget, ships quickly and is not fast fashion.
For a while, the assumption was that this spelled trouble for marketplaces. If your AI can shop the entire internet for you, who needs a marketplace at all?
Turns out, almost everyone.
A PSE Consulting survey of 4,250 AI-using shoppers across the UK, US, France and Germany found that 90 percent expect to use marketplaces the same amount or more as AI adoption grows. People are not abandoning marketplaces. They are splitting their behaviour into two parts. They want to use AI for discovery, to help them find the right thing. And they want a marketplace they trust for execution, to handle the payment, the delivery, and the quiet reassurance that if something goes wrong, someone has it covered.
Discovery and execution. Two separate jobs. When I read that line, I did smile, because for us, at Grimmor, that is not a prediction. It is the architecture behind our App.
What we actually built
Grimmor never had to choose a side, because we were built across both from day one.
The discovery layer is the AI feed and virtual try-on. The feed learns your taste and surfaces what actually suits you instead of drowning you in everything. The try-on lets you see a piece on your own body before you commit, which is the part that turns “maybe” into “yes.” That is the half people now want AI to handle.
The execution layer is the part that happens in Grimmor, without bouncing you to a checkout somewhere else: the brands, the payment, the delivery, all in one place. That is the half people still want a trusted marketplace to hold.
Most products in this space do one or the other. We do both, in a single scroll.
The objection I keep hearing
There is one stat in the same research that I keep hearing. 74 percent of AI shoppers say they would rather use an independent assistant like ChatGPT, and only 10 percent want an AI assistant built into an online store. On the surface, that sounds like bad news for anyone building AI into a shopping app.
So read it again, and notice the word: a store.
A single retailer’s chatbot is trying to sell you that one retailer’s shelf. Of course people do not trust it. It is conflicted by design. Grimmor’s AI is not a salesperson for one label. It is a discovery engine across hundreds of brands, which is exactly the role shoppers say they want AI to play: neutral, broad, on their side. And there is one thing an outside assistant cannot match: ChatGPT does not know your taste. Grimmor does. It learns what actually suits you, so it is neutral and personal at the same time. The trust they reserve for the marketplace is for the execution, and that trust is precisely what we hold. The stat that looks like an argument against us is actually an argument for what we are.
To Sum Up
The future of fashion commerce is not AI versus marketplaces. It is: AI-native marketplaces. The reports keep arriving at that conclusion in their own words, and every time, they are describing the same two things we set out to build: intelligent discovery and trusted execution, together, in one place.
The industry is still writing about where shopping is going. We have been living there for a while.
