KEYNOTE INSIGHT · 15 FEB 2026

THE AI-NATIVE BRAND: FASHION COMPANIES BUILT AROUND AI FROM THE START

Pau Almar argues that AI-native fashion brands, built around AI from the start rather than adding it as a tool, will repeat the challenge DNVBs posed in the 2010s, with a shorter window to respond. They will win first on personalisation and responsiveness. Incumbents should start with a unified, real-time customer data layer and AI-fluent commercial talent.

The fashion industry has discussed artificial intelligence for several years mainly as a set of tools. AI enters the conversation as a tool for trend forecasting, as a personalisation engine for e-commerce platforms, as a way to generate campaign imagery faster and cheaper. Each of these applications is real. A different question gets less attention.

The interesting question is what a fashion brand looks like when it is built around AI from the start — not retrofitted with it, not experimenting with it in a single department, but constituted by it in the same way that Digitally Native Vertical Brands of the 2010s were constituted by their relationship with data and direct-to-consumer channels.

Early versions of that brand already exist. I expect several to be significant commercial players within five years.

What the DNVB period showed

When Warby Parker launched in 2010, the fashion and retail establishment was reasonably polite about it. A nice DTC glasses brand. Disruptive for eyewear, perhaps. Not particularly threatening to the broader industry.

By 2015 it was clear that DNVBs — brands like Glossier, Everlane, Allbirds — represented a different business model, not only a different channel. They were built around data relationships with customers that legacy brands had delegated to wholesale partners and department stores for decades. They could test, iterate, and learn faster than traditional competitors. Their customer acquisition costs were lower, their feedback loops were tighter, their product development was more responsive.

The fashion establishment had roughly a decade to study this and respond. Most brands did something — built DTC channels, invested in CRM, appointed Chief Digital Officers. But the response was mostly additive: grafting digital capability onto existing organisations rather than rebuilding around the new logic.

AI-Native Vertical Brands will represent the next version of this challenge. The window to respond will be shorter.

The Distinction That Matters

A brand that uses AI is not an AI-native brand. This distinction sounds obvious stated plainly, but it is easy to miss in practice.

Zara uses AI tools across its operations. Mango, H&M, and every serious fashion company above a certain scale have invested in machine learning for demand forecasting, customer segmentation, and inventory optimisation. This is now standard. It does not constitute a different kind of company.

An AI-native brand is one where AI is not a capability added to existing processes — it is the process. Design is not aided by generative tools; design is a human-AI collaboration from the first sketch. Customer communication is not personalised by an algorithm running on a marketing platform; it is generated individually, in real time, based on context the AI has built from every prior interaction. Supply chain is not optimised by predictive tools; it is structured around AI-determined demand signals, with human intervention reserved for exceptions.

The operational implications of this distinction are significant. A traditional brand uses AI to make existing processes more efficient; an AI-native brand runs different processes.

What changes in practice

The most concrete expression of what changes is in personalisation — but not in the sense the industry has been discussing.

Current personalisation in fashion is largely segmentation: the algorithm places you in a cohort and shows you products popular with people like you. It is statistical and retrospective. It works, incrementally. It is not what the next generation of brands will do.

AI at the foundation means personalisation at the individual level, in real time, based on a continuously updated model of this specific customer — not a proxy group she resembles. The shopping journey she experiences is not a variation on a template; it is assembled for her, now, based on what she looked at last time, what she bought last season, what the weather is in her city this week, and a hundred other signals her behaviour has generated.

Stitch Fix has operated on this logic for years, with human stylists as the decision layer. The new generation removes the human bottleneck and extends the logic to the full brand experience — not just product selection but visual presentation, content, pricing offers, customer service.

The second material change is in production economics. One of the structural problems in fashion is the distance between when buying decisions are made and when market response can be measured. Traditional fashion brands commit to significant inventory months before the selling season, based on forecasts that are genuinely difficult to make accurately. The result, industry-wide, is enormous overproduction — an estimated 30% of all fashion produced globally is never sold at full price.

AI-native brands can compress this cycle in ways that change the economics. Not by being faster at the existing process, but by shifting more of the decision-making to a point when more information is available. Smaller initial production runs, based on AI-generated demand models, with rapid replenishment triggered by real sales data — and eventually, on-demand manufacturing for certain categories where customisation and speed can be delivered at a unit cost that makes economic sense.

The Threat to Incumbents Is Not Obvious

Large fashion brands looking at this shift tend to underestimate the threat for a reason that is structurally familiar: the threat is most dangerous to the business model that is most profitable right now.

Inditex, for example, has a large and efficient commercial operation. Its scale, its supply chain, its brand portfolio, its real estate — these are competitive advantages that took decades to build and are genuinely hard to replicate. But they are optimised for a specific way of doing business: large volume, managed risk, efficient distribution, brand presence in physical locations.

An AI-native brand entering the market is not competing for the same customer in the same way. It is building a direct, data-rich relationship with a specific segment of the market, iterating its product and experience in real time, and eliminating the categories of cost — wholesale margins, physical distribution overhead, large-batch production risk — that incumbents have learned to manage but not to remove.

The historical pattern here is consistent: the new entrant doesn't attack the incumbent's strongest position. It serves the customers the incumbent is not structured to serve optimally, builds capability, and then expands. DNVBs did not beat H&M in volume. They took the segment of the market that valued relationship and authenticity over price and convenience, and built durable businesses in that territory.

AI-native brands will do something similar. They will win in personalisation and responsiveness first, in the customer segments where those qualities matter most. Then they will build scale.

The Right Response

The right response for incumbent fashion brands is not to panic and it is not to wait. It is to assess where AI as foundation — not AI as tool — would most change the calculus of their specific business, and to move on those areas with real investment.

For most fashion brands, the first step is the customer data system. AI-native capabilities require a depth of customer data and the systems to act on it in real time. Most fashion companies have this data in fragmented form — CRM systems, e-commerce platforms, loyalty programmes that don't talk to each other cleanly. The structural prerequisite for AI-native operation is a unified, real-time customer data layer. Everything else depends on it.

The second requirement is talent. AI-native operation requires people who think about product, customer, and operations differently. Not data scientists grafted onto existing commercial teams, but commercial operators who are fluent in what AI can and cannot do — who think about brand and business strategy through the lens of what becomes possible when AI is at the centre.

The companies that handle this transition well will be the ones that understand what is actually changing, are honest about what their current organisation is and isn't built for, and make deliberate choices about which parts of the new logic to build and which to partner for.

// QUESTIONS

What is an AI-native fashion brand?

A brand where AI is the process rather than a capability added to existing processes. Design, customer communication and supply chain are built around AI from the start, with human intervention in the supply chain reserved for exceptions.

How will AI-native brands compete with Inditex and other fashion incumbents?

Pau Almar expects them to serve customers incumbents are not structured to serve, winning first in personalisation and responsiveness in the segments where those qualities matter most, then building scale, as DNVBs did in their own segment.

What should fashion brands do first to prepare for AI-native competitors?

Build a unified, real-time customer data layer, since most fashion companies hold customer data in fragmented CRM, e-commerce and loyalty systems. The second requirement is commercial operators fluent in what AI can and cannot do.

Companies in this piece: H&M, Inditex

← Back to Publications