AI is everywhere at NYFW 2027:
Front row, backstage and increasingly, in your shopping cart.
Alice + Olivia is using AI-powered virtual try-on, Google is helping designers Jane Wade and Sergio Hudson visualize collections and runway staging before production, and Meta’s new AI agent Muse can actually shop for you.
So we're getting remarkably good at answering:
Will I like it?
What will it look like on me?
Can you buy it for me?
There's just one fairly important question left:
Will it fit? And will it fit the way I want it to fit?
Because before AI can revolutionize how we shop, fashion has a pretty basic infrastructure problem to solve.
Sizing.
We're building an intelligence layer on top of a data layer that's basically vibes.
And Fashion Week itself is inadvertently demonstrating the gap.
Jane Wade's new AI tools can help her visualize a garment before it's produced. But she still says they don't replace live fittings.
Virtual try-on can generate an increasingly convincing image of you wearing a dress. But visualization isn't measurement. FashLogic, another AI virtual try-on platform, spells this out unusually clearly in its own FAQ: its technology can show silhouette, color and proportion, but it can't tell you which size will fit. For that, you're sent back to the retailer's size chart.
AI can see me in the clothes.
AI can buy me the clothes.
It still needs fashion to tell it how big the clothes actually are.
Here's the maddening part: we aren't starting from zero.
The ISO 8559 family of standards gives the fashion industry a standardized vocabulary for measuring human bodies and designating clothing sizes. Its roots go back decades, and the current standards define measurements that can even be used to build physical and digital anthropometric databases.
But the standard is based on body measurements, not garment measurements.
And that's the missing half of the equation.
Knowing that my waist is X centimeters is useful. Knowing that a pair of trousers has a waistband of Y centimeters, a particular rise, hip measurement, inseam, leg opening and amount of stretch is what allows a machine to reason about how those trousers will actually fit me.
Today, that information is inconsistent, incomplete, buried in size charts or simply nonexistent.
So an AI can know your exact measurements. Your proportions. Your style. How you like things to fit. What's already in your closet.
But if the product data gives it S, M or L?
Your AI is guessing.
So is the agent that's supposed to be shopping for you.
One of the more interesting new fashion-tech companies I've seen this week is Virtue Mirage, precisely because it's starting to attack this problem from the measurement side rather than just the image side.
Its “digital twin” can be created from photos or actual body measurements and carries those measurements across participating stores. Its size advisor then combines the shopper's measurements with a brand's size chart to recommend a size.
That's much closer to where this needs to go.
But even the perfect digital twin of me only solves half the problem.
We need a digital twin of the clothes.
Imagine if every garment online came with standardized, machine-readable dimensions. Not just “M,” but the measurements and attributes that determine how that M actually fits.
Then the agent gets interesting.
Not:
“Find me some wide-leg trousers.”
But:
“Find me trousers that fit at my waist, give me this much room through my hips, fall this far from the floor and have the same silhouette as the pair I already love.”
And it could compare those preferences against actual product data across thousands of garments.
This isn't just a consumer problem, either.
At COTERIE New York this week, Centric Software was making almost the same infrastructure argument from the other side of the industry. Its pitch for AI-powered product development is built around connecting AI directly to live product data across the fashion lifecycle. The company called connected product data “the foundation for AI.”
Exactly.
Fashion is starting to understand that AI is only as useful as the data underneath it. Now that thinking needs to make it all the way to the product page.
The real promise of AI isn't a stylist telling me what looks cute. It's an agent that knows what I like, knows what fits and can actually buy it for me.
Before fashion gives AI a seat at the runway, maybe give it a tape measure.
Solely,
Lucy