Why a pair of shorts follows you online, and how KinraLabs plans to fix fashion's keyhole view of shoppers

"I hired a stylist myself because I wanted to figure out what I should wear and how I should dress, and I hated the experience so deeply that I decided I needed to do it differently," said Olga, founder and CEO of KinraLabs.

"She was putting me into categories: your body is this, your colour is this. I don't believe in that," Olga said of the stylist, who told her to avoid red and to wear skinny jeans when she preferred baggy ones. "I believe fashion is about self-expression, and I wanted my personality to be expressed through how I dress, rather than my body."

About five years ago she took a styling course, then spent the years that followed working with private clients and developing a methodology of her own, built around each client's personality, habits and sense of self. "My clients would tell me it's a bit like a therapy session, and I'd say, 'Yes, a little bit,'" she said. "I need to know who you are to be able to dress you."

That practice sits alongside more than a decade in consulting, where Olga specialised in customer experience management and helped some of the world's largest corporations, retailers among them, redesign how they serve customers. KinraLabs is where the two careers meet: an AI platform that builds a living style profile owned by the shopper, which retailers and AI shopping assistants can draw on.

Retail personalisation still sees the shopper through a keyhole

"Think about what fashion retail knows about you today: what you clicked, what you browsed, sometimes what you bought, maybe what you returned," Olga said. "It's a very narrow view of who you are, like looking through a keyhole, and retailers are spending billions trying to optimise within that view."

Her example is one most online shoppers will recognise: a pair of blue shorts viewed once, perhaps by mistake, perhaps as a gift, perhaps before buying them elsewhere, which then follow the shopper around the internet for a month. "The systems today take that single action as a signal that you really need it," she said. "In fashion, the only way to truly personalise is to broaden that view. You need to see what's behind the door."

"You might have bought something and never touched it, and it's been sitting in your wardrobe for a year. Retailers don't have access to any of those signals," she added. KinraLabs pairs a customer's identity with real-life signals about what they wear, like and dislike to build a profile that evolves over time. "The key differentiator is that it's a customer profile that you own as a customer, and retailers plug into it."

That ownership model has implications well beyond fashion. Personalisation across e-commerce has largely been built on behavioural data held by retailers, and a portable profile held by the customer shifts control over the context behind every recommendation.

Five years of private clients served as the first proof of concept

"I treat my five years of styling work as a concierge MVP," Olga said. "I validated it with clients one to one, and I also saw the commercial impact it makes." Clients who received a curated selection with an explanation of why each item suited them bought more, returned less and went back to the same brands, she said. "Those brands now fit their identity."

"I haven't seen a major difference in behaviour," she said of the vibe-coded prototype she built next, without the full technology behind it, to test whether people would respond to a digital experience as they did to her in person. "There's much more to do, but in terms of trends and dimensions it was still relevant." Conversations with independent brands and large e-commerce players followed to test positioning before she built the production platform, which combines a multimodal AI model with a proprietary machine learning model.

AI shopping assistants forced a rethink of who the product serves

"There have been more pivots than a straight line," Olga said. The most significant came just as she had settled on a value proposition for retailers, built around direct integration or a white-label product, and AI shopping assistants began to appear. "I thought, 'Interesting. Now I need to pivot,'" she said. "How do I package the solution and build the technology so that AI shopping assistants can plug into it as well?"

"You might ask an AI shopping assistant to find you a summer dress that matches your style," she explained. "Right now the protocols are optimised for the retailer's catalogue. They don't define what your style is, and they don't have a clean data source for what that means." Her response is to position KinraLabs as a style layer those assistants can query. As agentic commerce moves from demonstration into real purchases, the question of whose data describes the shopper, and who owns it, is likely to decide which companies capture value from AI-mediated retail.

Independent labels and large retailers buy two different products

"A lot of people don't know the local brands, and it's very hard to find them, because the algorithms don't empower the visibility of small brands," Olga said. Independent brands, with an emphasis on regional labels, are integrated directly into the KinraLabs platform, where a shopper identifies her style and sees items matched to it. A few are already live. "Fashion is very fragmented. We can have multiple brands on us at the same time," she said. For small labels, appearing in front of the right customer brings discovery, higher purchase intent and loyalty.

"It becomes part of their tech stack, fully built into how customers experience their own platform," she said of the white-label integration offered to larger single-brand and multi-brand retailers, which carry enough stock to serve a customer on their own. That route requires a larger technology investment from retailers, and KinraLabs is currently running proofs of concept that need no integration effort, so retailers can gain confidence before committing to a full technical integration.

Market estimates run wide, and the consumer offer covers womenswear for now

"That alone within the region is a 5 billion market, just on simple bottom-up maths," Olga said of her estimate, built from the number of independent brands, larger brands with sufficient selection and full-scale e-commerce platforms. Figures for annual spend on personalisation software vary sharply, from 12 billion to 50 billion by her reading. "I usually pick the lowest, because that's already huge enough."

"The main unifying factor is women who actually want to solve that problem. They feel it, and they want to find a solution," she said of a consumer base that currently covers womenswear only, with menswear next, and that has drawn engagement across generations. The company's limits at this stage are clear: a handful of independent brands are live, the larger retail business remains at proof-of-concept stage, and positioning has already been reworked once in response to AI shopping assistants.

Ambitions extend beyond the UAE and beyond clothing

"We start within the UAE, grow regionally and then expand globally," Olga said. "I believe it will be global."

"Style isn't just about the clothes we wear. It's also makeup and fragrance," she added. For the shopper still being followed around the internet by a pair of shorts she never wanted, the pitch is a recommendation engine that knows who she is. Whether retailers and AI assistants agree to plug into a profile they do not own will determine how far that ambition travels.