ChatGPT Now Tries Clothes On for You. Store Design That Still Hides the Fitting Room Is Dead

On 1 October, OpenAI added a new button to ChatGPT: “Try On”. You upload a selfie or a full-body photo, pick a garment and the machine puts it on you. No queue, no curtain that won’t close, no overhead light that makes you look freshly discharged from intensive care. And while half the industry cheers or panics, I have a different question: if a chatbot can try clothes on you from your sofa, what excuse does your fitting room have for still being the worst square metre in the shop? Because this looks like a tech story, but it’s a store design problem. And we’ve been failing it for years.

According to TechCrunch, the feature runs on OpenAI’s new image model, promises more natural lighting and richer textures, and comes with a Favorites library where you save products next to your virtual try-ons. It isn’t the first: Google launched its own virtual try-on in 2025. But when the assistant hundreds of millions of people use for everything starts dressing them too, the message is clear. Trying on clothes is moving house. And it isn’t moving into your store.

AI takes the image. The store keeps the body

Let’s be precise, because there’s a lot of nonsense flying around. A virtual try-on answers one question: “how does this look on me in a picture?”. And it answers it better every month. What it doesn’t answer is everything else. Whether it itches. Whether the sleeve pulls when you lift your arm. Whether the trousers survive sitting down. Whether that green, under your office lights, drains your face. Whether this brand’s medium is the other brand’s medium.

That’s the hole, and it’s an expensive one. A late-2024 study by McKinsey and The Business of Fashion, cited by Fortune, estimates that up to 30% of fashion bought online is sent back, and each return costs on average between $21 and $46 in shipping, handling and making the item sellable again. AI may bring that number down. I hope it does. But the doubt behind most returns isn’t visual. It’s physical.

You can’t render touch

You can generate the perfect image of a sweatshirt. You can’t generate the gesture of pinching the fabric between two fingers and knowing, in half a second, whether it’s good or cardboard. That gesture is physical retail’s oldest competitive advantage and its most wasted one. Every time a store hides product behind a counter, piles it up with no order or hangs it so tight nobody can touch it, it’s giving away its only trump card against the screen.

Your fitting room: where the sale is decided, designed like a storeroom

Now for the part that hurts. Walk into your store as a customer and head straight for the fitting room. What do you find? I know it by heart, because I’ve seen it in chains and in neighbourhood shops alike: a curtain that doesn’t quite close, one hook for one hanger, a stool that can’t hold a coat, a mirror that distorts and a light that turns anyone into a ghost. And sometimes, as a bonus, boxes of stock in the corridor.

It’s absurd. The fitting room is the only place in the store where the customer has already chosen a product and is one step away from paying. Years ago, RetailNext data published in Retail TouchPoints already showed that fitting room visitors were 25% more likely to buy. And a UNC Kenan-Flagler study cited in the same article found that adding just one staff member to the fitting room area raised transactions per hour by 12.55%. One person. Not a screen. Not an app.

What a fitting room needs to compete with AI

None of this requires a refit. Warm, frontal light that flatters instead of punishing. A full-length mirror with no distortion, and another to see your back. Room to put down a bag and a coat without a balancing act. Hooks for three or four pieces, because nobody tries on just one. And someone who asks whether you need another size before you have to get dressed again to go and find it. That’s real customer experience, not the slide-deck kind.

Luxury worked this out long ago. In its boutiques, the area where you choose and try on isn’t a corridor with curtains: it’s a salon. Few pieces, a comfortable seat, kind light, time. It isn’t decoration. It’s a way of telling the customer their decision matters.

“AI will kill fitting rooms, so why invest?”

That’s the objection I can already hear. And it has spreadsheet logic: if trying on happens on the phone, the fitting room is a cost. Cut the square metres, close cubicles, let customers try at home and return if they must.

Wrong. AI won’t kill the physical try-on; it will fill it with more decided customers. Someone who walks into your store after seeing themselves in that jacket on ChatGPT isn’t browsing. They’re coming to confirm. They bring a half-made decision and all that’s missing is the body: touching, moving, seeing themselves in real light. If your fitting room lets them down at that exact moment, you don’t just lose the sale. You prove the screen right. AI is the best generator of qualified traffic your fitting room will ever have. The phygital customer journey starts on the phone and closes, or breaks, behind your curtain.

One more thing: the day virtual try-ons work beautifully, the store that survives won’t be the one with the most square metres. It’ll be the one offering what the algorithm can’t: material, fit and someone who knows the product.

Self-criticism: we visual merchandisers spend our talent on the glass

Now it’s my turn, and my trade’s. Those of us in visual merchandising and window display have been accomplices in this neglect. We love the window: it’s visible, photogenic, it makes the magazines and our portfolios. Nobody photographs a fitting room. So it gets the leftover budget, the leftover design and the leftover attention. In other words, none.

We build signature installations behind the glass to pull people in, then send them to a cubicle that contradicts everything the window promised. That isn’t retail design: it’s false advertising with hangers. Serious store design has to treat the fitting room as the end of a story that begins on the pavement, with the same care, the same light and the same intention. If we don’t do it, a chatbot will.

Make your fitting room beat the screen

The question is no longer whether AI will steal your customers. It’s what the ones who still come will find. Start with the cheap stuff: count how many people go into the fitting room and how many come out with something in hand, change the light, clear the corridor, put someone attentive there. Then think of the fitting room as what it is: the final scene of your store.

If you want someone to look at it through a customer’s eyes, the Phygital Retail Audit from my Consultancy reviews your window, customer flow, fitting rooms and website in one on-site day, with a prioritised report. If you already know it’s time to redesign, at the BARRONVISUAL Studio we design stores from the pavement to the fitting room. And if you want to learn to do it with your team, there’s the Academy. Not sure? Book a 20-minute Diagnosis with me.

Now it’s your turn: when was the last time you tried something on in your own store? Tell me in the comments.

Frequently asked questions

What is ChatGPT’s virtual try-on?

It’s a feature OpenAI launched on 1 October 2026: you upload a selfie or full-body photo and ChatGPT generates an image of you wearing the chosen garment or accessory. It works from shopping results or from a product image.

Will virtual try-ons replace store fitting rooms?

They solve the visual doubt, but not touch, fit or comfort. Most likely they’ll send more decided customers to stores, who will expect a fitting room good enough to confirm the purchase.

How can I improve my store’s fitting rooms without a refit?

Switch to warm, frontal lighting, use full-length mirrors without distortion, add hooks and a seat, clear the corridor and assign someone to the area. These are low-cost changes with a direct impact on conversion rate.

Which store KPIs show whether a fitting room works?

How many people enter the fitting room, how many items they try on and how many leave with a purchase. Comparing them before and after each change tells you whether the fitting room design is helping or holding back sales.

---

Sources (accessed October 2026): TechCrunch, ChatGPT virtual try-on · TechCrunch, Google virtual try-on (2025) · Fortune, McKinsey and The Business of Fashion returns data · Retail TouchPoints, RetailNext and UNC Kenan-Flagler study