ASOS, the British retailer behind multiple in-house brands and Topshop, has been one of the driving forces behind the fast-fashion revolution. In this conversation with McKinsey’s Hai-Ly Nguyen and Holger Harreis, the company’s chief technology officer Przemek Czarnecki discusses the impact of AI on ASOS’s operations, which sell hundreds of partner brands and its own labels to consumers in more than 150 countries. The interview has been edited for clarity and length.
The most significant challenge in scaling AI is not technological but organizational. The biggest mistake a company can make is investing in use cases that are not commercially meaningful.
Q: How is AI reshaping ASOS’s competitive advantage as a digital-first fashion retailer?
Przemek Czarnecki: AI is fundamentally reshaping our competitive advantage across three key dimensions. First, it allows us to understand customer intent at a much deeper level, moving beyond basic keywords to interpret even vague, conversational search queries. Second, it powers a new frontier of hyperpersonalization, where we can generate highly relevant content and recommendations tailored to the individual. Third, it drives significant productivity improvements, most evident in our customer care, where AI agents now handle 50 percent of inbound requests. Since the underlying technology is not proprietary, our true competitive advantage comes from the speed and quality of deployment—how effectively we integrate these tools to deliver a superior experience.
Q: Where are you seeing the biggest tangible impact from AI today?
Czarnecki: We approach our AI deployment through a phased road map designed to ensure a clear return on investment. We started first with software development—phase zero. Then, in phase one, we focused on our call center, tackling the ten most-repetitive customer inquiries and making our service operations extremely productive. Concurrently, in phase one, we democratized foundational AI by rolling out general Copilot tools to our entire workforce, achieving an exceptional 90 percent adoption rate.
Phase two is deploying targeted back-office agents across HR, legal, and finance to autonomously handle internal policies and IT help desks. Running slightly after but mostly in parallel is phase three, which applies agentic AI to the very core of our fashion business—buying, design, and merchandising—enabling rapid, machine-driven experimentation. Ultimately, our overarching vision is to build a “hybrid organization” where the corporate calendar is populated by both human employees and autonomous agents working seamlessly together to drive commercial value.
Q: What are the most critical enablers that make or break a successful AI transformation?
Czarnecki: Successful value creation from AI depends on several foundational enablers. The most critical is the quality and accessibility of data and content; without this, even the most advanced models will fail. Moreover, for AI agents to be truly impactful, they must be empowered to take action, which requires a robust and sufficient layer of APIs across business systems. Finally, success hinges on having the right talent and partners. This involves not only cultivating internal skills but also making strategic choices about which providers and technology partners to work with, because the capabilities between them can differ significantly.
Q: What has been the hardest part of scaling AI across ASOS?
Czarnecki: The biggest mistake a company can make is investing in use cases that are not commercially meaningful. This leads to a fragmentation of investments into low-impact projects that, while technically successful, fail to deliver strong returns and never truly scale. The hardest part is therefore establishing a strong operating model and the strategic discipline to identify and prioritize only the highest-value use cases. This prevents the proliferation of “nice to have” agents and ensures development capacity is focused on initiatives that move the needle.
Q: How is AI changing the way your teams work and make decisions as well as the skill requirements of your workforce?
Czarnecki: The workforce transformation agenda at ASOS is structured around three capability pillars. The first is data fluency: equipping employees across functions with the ability to interrogate data and generate actionable insights—a skill set that, while widely assumed to be present, frequently requires deliberate reinforcement. The second is the democratization of agent creation, enabling employees to build their own low-code or no-code agents and self-serve automation without dependence on a centralized engineering resource. The third and most strategically differentiated is the cultivation of an internal AI strategist capability—individuals who can identify high-value agentic use cases, redesign processes around them, and ensure AI investments translate into measurable impact rather than technically sound but marginal initiatives. Without this function, organizations risk a proliferation of low-impact use cases that consume capacity without delivering material return.
Q: How do you see AI—particularly AI-driven discovery and shopping agents—reshaping online fashion retail, and how is ASOS preparing for that shift?
Czarnecki: The rise of agentic commerce represents a structural shift in how consumers will engage with retail, and ASOS’s strategic response is anchored in a single, clear guiding principle: Agentic shoppers and human shoppers must be treated equally well.
