We took a close look at Stylique Technologies, the plug-in AI layer that lets shoppers on an online fashion store see an item on themselves and get a size call before checkout. Inside: why fashion's returns crisis is a confidence problem in a logistics costume, which merchants plug-in try-on actually wins with, and the number this whole category keeps measuring wrong.
Take an apparel brand running 10-20% net margins. Now hand it a 30-35% return rate driven mostly by sizing uncertainty, and watch returns processing eat what's left. That is the actual day job at most online fashion stores: traffic arrives fine, conversion sits stuck between 0.5% and 2%, and the checkout queue is full of shoppers quietly asking three questions. Will this fit me? Is this cut tight or relaxed? What happens if I need to send it back? Some order two sizes and plan the return before the parcel ships. Many just leave. Stylique's starting claim is that better photos, size charts, and fit notes all fail the same way: they ask the shopper to imagine. The company puts it bluntly. The assumption that customers can accurately picture a garment on their own body from a product photo is fundamentally broken. They can't.
The wedge: one confidence layer, not another size chart
The product is a 2D, photo-based try-on. A shopper uploads a picture, sees the garment on themselves, and gets an instant size recommendation plus style and outfit suggestions. No 3D assets to build. No body scan. It is mobile-first and plugs into a Shopify-style storefront instead of demanding a rebuild.
The hard part to copy is not the model, because 2D garment generation is commoditising fast. The wedge is the packaging. Try-on, sizing, styling, and analytics ship as one layer rather than three point tools and a dashboard nobody opens, and each piece feeds the next: try-on behaviour sharpens the size call, and sizing data tells a brand which SKUs run small before the returns do. Pricing follows the same logic, metered by traffic and volume: 1,000 try-ons a month at Starter under 25,000 visitors, 3,500 at Growth, 7,500 plus a dedicated success manager at Enterprise. That maps to how a small brand actually grows, not how an enterprise procurement team buys.
One detail deserves more credit than it gets: privacy-first image handling with no human access. Photo upload is the biggest drop-off risk in any try-on flow, since shoppers will not send a body shot to a boutique if they suspect a person might see it.
Now the caveat, stated plainly. Every number is self-reported. Fifty-plus brands, returns down 25%, conversion up 20%, 80,000+ engagements, AOV up 20%, with stated outcome ranges of 15-25% more add-to-cart and 10-15% higher order value. No logos. No named case studies yet. Directionally consistent with what this category tends to produce, but it is platform data, not proof.
The ICP they actually win: the brands Nike's vendors never called
The buyer is a founder, an ecommerce manager, or a head of digital, and the sale runs through a demo rather than a procurement cycle. The sweet spot is sub-$10M GMV online apparel: Shopify merchants, boutiques, and labels with real traffic and a fit-shaped returns problem. The About page goes further and names small and medium retailers in developing countries who lack access to advanced AI, which is a sharper positioning move than it looks. Enterprise virtual try-on already exists at Nike, ASOS, Warby Parker, Zara, and Gucci (deployments Stylique cites as industry examples, not customers), and it ran on 3D pipelines and budgets the long tail will never see. Stylique's bet is that the identical decision problem shows up at 40,000 monthly visitors, and almost nobody is selling to it.
Two qualifiers. Brands whose returns skew toward colour, fabric quality, or buyer's remorse will see little movement, because this fixes fit. And stores under roughly 10,000 monthly visitors have little for the layer to amplify, since try-on converts existing intent rather than creating it.
What the category still gets wrong: it optimises the picture, not the decision
The status quo optimises production value. Enterprise 3D try-on optimises fidelity and spectacle, because spectacular demos sell six-figure contracts. Both measure engagement, dwell time, and press coverage. Wrong number. The number that pays rent is decision-stage conversion and returns per order, and a shopper does not need photorealism to answer the two questions that actually block purchase: will it fit, and does this silhouette work on my body. A 2D approximation answers both in seconds, on a phone.
The deeper misdiagnosis sits upstream. Brands read 0.5-2% conversion as a traffic problem and buy more ads to fill a bucket with a hole in the bottom. Then they read returns as a logistics problem and negotiate cheaper processing, when the return was decided before checkout ever happened. You cannot reverse-logistics your way out of a pre-purchase guess.
Engagement deserves suspicion too. Stylique reports 3x engagement, 55% more multi-product exploration, and 25-45% more time on product pages. Fine. Dwell time without decisions is entertainment. The metric to demand from anyone in this space, Stylique included, is add-to-cart on try-on sessions versus non-try-on sessions, and returns segmented by fit reason.
For operators watching this space, the tell will be second-season retention: brands that keep the layer through two peak periods are the real evidence, and that data does not exist yet for a company this early. If you run an apparel store, the cheapest evaluation is a session split, not a demo. Stylique's framing survives scrutiny even where its proof does not. They say they are not selling visuals, and they are right. They are selling the end of guessing, and in a category still obsessed with how the picture looks, that is the correct number to sell against.