All About AI: Evgeni Zapadinsky & Lana Vereshchagina of Aiclo

We’re observing a big disconnect (in both our own fashion-only data and in wider benchmarks from other sources) between the increasing capability and reliability of AI, in both a general sense and in discrete applications, and the lack of trust that end users place in its outputs. What does trust in your specific applications of AI look like, and how do you achieve and measure it?

Lana: When you look at the broader AI landscape, the lack of trust in fashion makes perfect sense. The industry has been overwhelmed by generative tools that produce stunning visuals, but from a manufacturing perspective, they’re essentially hallucinations. They don’t understand fabric physics or grading logic. In production, a mistake in a pattern isn’t just a bad pixel. It’s wasted fabric. It’s a financial liability.

For Aiclo, trust is defined by real-world outcomes. It means a pattern maker reclaims hours of their day because they’re not drafting a basic block from scratch anymore. It means a brand can launch an entire made-to-measure collection without sewing a single physical sample. And ultimately, it means the end consumer receives a garment that fits their body.

We achieve that trust by building a true parametric engine, not just a visual layer on top of something vague. But more importantly, trust comes from the ability to verify. We don’t just generate a file and ask for blind faith. We give professionals the tools to check the math before anything is physically cut. A technical designer can compare the AI-generated pattern measurements directly against the avatar’s body measurements, evaluate fit and strain maps, make structural adjustments based on specific fabric properties. We’re not asking the industry to adopt a closed black box. Once the pattern is verified inside Aiclo, it exports as standard production-ready DXF and PDF files that integrate into any factory’s existing CAD workflow.

We measure that trust in hard numbers: 14,500 fashion professionals already using our engine, and enterprise partners routing our files directly to their cutting rooms.

Personalisation is suddenly everywhere in fashion, but it’s primarily talked about as a tool for engaging shoppers in an individualised way — putting pre-existing products in front of them in a way that feels unique, and that maximises conversions. Product-level personalisation is a different beast entirely, when we scope it to look beyond the confines of garment decoration and artwork. What does it mean, from your perspective, for fashion to take a meaningful stab at making mass customisation commercially viable and operationally scalable as a business model?

Evgeni: There are a lot of studies showing that mass customisation isn’t just better for sustainability, it’s actually more economically efficient than mass production. Both sides win. Consumers get garments that are custom-designed and made to their measurements. Manufacturers produce only what’s been sold. No overstock, no guessing.

That said, and this is the part that doesn’t always make it into those studies, the transition for established brands is genuinely expensive and disruptive. It’s not a software update. It’s a rethinking of the entire production logic.

Aiclo’s role in that shift is very specific. We automate the design-to-pattern step: you get made-to-measure patterns in seconds, a tech pack ready for production, plus a virtual try-on. After that, all that’s left is to cut the fabric and sew it. But, honestly, even that “just cut and sew” step isn’t straightforward when you’re talking about large factories. They’re built for mass production runs. Reconfiguring them for custom, one-off garments is a completely different operational challenge.

So our bet is on a different path. We’re working with smaller manufacturers. The vision is that in a few years there’s a network of these companies, all running on Aiclo, producing bespoke apparel at scale. That network becomes the seed for a real transition from mass production to mass customisation. Not a mandate from the top, but an economic reality that proves itself.

There are two layers to made-to-measure and mass customisation: the part the shopper sees, and the infrastructure and production network that’s then needed to go and execute on it. It’s one thing to say that everyone should be able to get a garment that fits them uniquely, or that combines their individual aesthetic vision with the brand’s. It’s another to stand up and then scale up the upstream side of things. What investments do you believe need to be made there? And do you see those investments starting to materialise already?

Evgeni: Right, so there are two completely separate problems here. The consumer-facing side (the idea that everyone should be able to get something that fits them, that reflects their aesthetic), that part is actually solvable today. We solve it. The harder problem is everything that happens after someone places that order.

Every stage of the supply chain has a bottleneck right now. Design, presentation, manufacturing, delivery: there isn’t a single stage that’s fully ready for true custom, on-demand production at scale. Aiclo addresses the pattern engineering piece and can virtually present the garment on any individual. But the factory floor? That’s still a real constraint.

Honestly, what we’d love to see, and what we think would genuinely move the industry, is a manufacturer that commits to building a workshop around this model. Not a high-end atelier, we already have those. I mean a manufacturer of bespoke everyday clothing. If we were an established brand, that’s exactly where we’d place an experimental bet: set up a small production unit that runs on the made-to-measure, on-demand model, let it operate in parallel with the main line, see what breaks and what works. That kind of experiment doesn’t just prove a concept; if it succeeds, it gives the whole industry a replicable blueprint.

The investments need to happen at the infrastructure level. Flexible production lines, cutting infrastructure that handles individual units, logistics built for single-order fulfilment rather than bulk runs. Some of it is starting to materialise, but slowly. The economics haven’t yet forced the hand of the big players. When they do, and I think they will, the transition will be fast.

Your aim is to combine “the math of a master pattern maker” with “the speed of a neural network”. Implicit in that blend is the idea that AI should be used to provide something that humans can’t, and vice versa. How are you seeing that play out? Are pattern makers ready to have AI work alongside them? Are companies finding AI to be a viable way to release the pressure that’s being placed on scared, specialised resources?

Lana: The reality right now is that there’s a severe talent shortage in technical design. The specialists we have are often overwhelmed, frankly burning out from the sheer volume of repetitive math that basic drafting and grading requires.

There’s a common misconception that AI has to do something humans simply can’t. But that’s not really the point. What AI does is process vast amounts of data in ways that change what’s even noticeable. When you train on anthropometric data, neural networks surface correlations that a human would genuinely miss; not because they’re not smart enough, but because nobody’s looking at a neck measurement and simultaneously thinking about leg musculature on a 3D avatar. And beyond that, an algorithm doesn’t forget to add a seam allowance after an eight-hour shift. Basic human error just drops out of the equation.

Are pattern makers ready for this? Absolutely. A significant portion of our users are professional pattern makers. For years they watched AI disrupt creative fields (image generation, copywriting) while their highly technical, gruelling work was left completely untouched. They’ve been waiting for a tool built specifically for them. They don’t want to spend eight hours on a basic sleeve block. They want to get straight to the fitting.

For companies, this is genuinely the only viable way to relieve pressure on their teams. Adopting AI here doesn’t mean pattern makers lose their jobs. It means they become technical directors: overseeing the AI, focusing on high-level fit and drape, doing the work that actually requires human judgment. It multiplies their output without asking them to disappear.

In some senses, the conditions have never been better for technology startups and scale-ups: a step change in the technology itself creates a wide-open horizon of opportunity. In other senses, the market is more competitive and more hostile than ever, with a wide array of new companies vying for technology spend that can’t possibly cover every outcome. How do you see Aiclo’s market position developing over the next couple of years? What’s your vision for the project?

Evgeni: The conditions have never been better, and the market has never been noisier. Right now the vast majority of AI startups in fashion are focused on the visual side: mood boards, marketing imagery, virtual try-ons that look great in a deck but don’t connect to anything on the production side.

Aiclo’s position is strictly technical. We’re not building a visual toy. We’re training AI to solve engineering problems: automating the complex underlying geometry of clothing. That’s a narrower bet, but it’s the right one, because that’s where the actual cost sits.

Looking ahead: the vision is to become a seamless part of everyday workflow for technical designers, to the point where the tool disappears into the process. When technical designers stop worrying about the mathematical grind, they have the freedom to create more complex, better-fitting work. That’s the unlock. And the market will catch up, not because of enthusiasm for AI, but because the economics of manual drafting simply become indefensible at scale.

And while Aiclo is built on serious engineering, it doesn’t require technical expertise to use. DIY sewers and fashion enthusiasts can design a garment they have in mind, see it on their own body, and share the result with friends or on social networks. The fit is generated automatically from individual measurements, and so are the patterns. So we believe that in a few years Aiclo will meet the demands of a wide audience, from technical designers and business units all the way up to fashion enthusiasts.

Last year, we asked technology executives whether they predicted AI would become more obvious, as a primary interface paradigm, or whether it would disappear into the background in the way that vital but invisible platforms like AWS and Azure have done. This year’s data suggests that the answer is both: fashion professionals that use AI interact with it as both the engine and the steering wheel. But we also see that, for a lot of companies, both kinds of applications are still largely in the scoping or refined-pilot stage. What do you see being the trigger for deeper adoption and wider roll-out?

Lana: The reason so many companies are stuck in the pilot stage isn’t lack of interest. It’s that they’ve been looking for value in the wrong places. For years, “AI in fashion” meant either high-gloss marketing tools or isolated IT experiments that never bridged to the factory floor. Neither one touches the core of the product creation process.

The trigger for mass adoption will be the shift from “AI for show” to “AI for unit economics.” Manual drafting and grading are major cost centers that kill the profitability of on-demand models. An API-driven, sell-then-make workflow is the only way to reclaim those hours and eliminate the cost of physical sampling.

Evgeni: Beyond the economics, the other trigger is physical. Industry 4.0, robotics, advanced cutting machines. As manufacturing becomes more automated, robotic sewing lines need a continuous feed of precise, machine-readable digital patterns. There’s no room for manual intervention in that loop. And that’s exactly what Aiclo is built for: it reduces cost, compresses the production cycle, and is designed from the ground up to feed automated manufacturing lines with production-ready files.

Adoption deepens when AI stops being something you log into and becomes a silent engine that translates a sale into a production-ready pattern. When mass customisation stops being a niche project and becomes the operational default, the model that makes physical sampling and overstock structurally obsolete, that’s when it becomes the backbone of the industry. Not a trend. The default.