Dispelling the magic: 5 retail realities about AI in fashion | Mark Lewis
Fashion retailers need less AI hype and more clarity. Lifetime retailer and technological strategist Mark Lewis, knows this better than anyone else. Having directed IT teams and having implemented solutions of all types, he has a clear understanding of the need to “demystify and de-fearify” (to borrow his phrase) the reality of AI, how it works in practice and how merchandisers can use it to meet their main objectives.
As scientist and science writer Arthur C. Clarke famously said, “Any sufficiently advanced technology is indistinguishable from magic.” But in retail, an AI-as-magic mindset may be holding retailers back if it causes them to fear or distrust it.
It’s time to pull back the curtain and explain the reality of AI and how it can be materially different, and better, than how most fashion retailers are currently managing operations.
Reality #1: Behind AI’s ‘magic’ is maths, not mystery
For many fashion retailers, AI still feels mysterious, overpromised or even intimidating. The industry is full of grand claims and “AI-powered” labels. Yet many retail leaders wonder how it’s different from what they’ve been doing for decades.
As Mark Lewis puts it, “AI isn’t magic, it’s logic, data and testing, applied at scale”. In other words, it’s sophisticated maths. In inventory management and optimization, these maths can calculate up-to-the-minute demand through millions of backtests on product attributes like material, sub-brand or heel height to uncover which combinations of locations truly drive sales.
Not understanding what AI actually is (and isn’t) keeps retailers stuck making highly complex decisions in spreadsheets and with gut feel. Meanwhile, practical, high-impact gains like smarter demand forecasting and inventory optimization are left on the table.
Reality #2: The biggest AI barrier in fashion is mindset, not tech
It would be an oversimplification to say hesitation around AI is purely technical. It’s also cultural and driven by a lack of clarity, perceived complexity or fear of losing control.
Many fashion retailers do see the need to evolve but have been burned before. As Lewis notes “Most aren’t resisting AI itself, they’re resisting what they fear will be another complex, time-draining, resource-hungry project.” After decades of ERP implementations that promised transformation but delivered fatigue, it’s understandable. “You’ve got fires everywhere, cash is tight, teams are stretched,” Lewis says. “Someone comes along offering AI and you think: I just don’t have the brain space for this.”
But that resistance has a cost. Retailers who embrace AI intelligently are pulling ahead because they’re reacting faster, holding less stock, and staying relevant to customers. The success of these retailers is their responsiveness to demand versus slower competitors who can’t even see it coming.
Reality #3: The most valuable AI isn’t flashy, it’s practical
While the spotlight often falls on triendier applications like design or virtual models, the highest, quickest value in fashion comes from practical use cases.
Take inventory optimization. Merchandising has been traditionally “push-based” in which stock is pushed from the warehouse based on sales forecasts and category plans leading to expensive inter-branch transfers. But as Lewis explains, “There’s no room for error anymore, and no time to get it wrong and fix it later because at that point, you’ve lost the customer.”
AI enables a “pull-based” approach. Instead of stock being pushed out based on assumption, it’s pulled toward real demand from customers. With the ability to constantly regenerate and reevaluate where the right stock should be and when, retailers can forecast with remarkable precision, even in highly-fragmented markets.
When push comes to shove, AI lets fashion retailers pull instead.
Reality #4: Some fashion retailers have been using AI for years
Another common misconception is that all AI solutions are new and that AI itself is just one thing.
In fact, as Lewis points out, “everyone says they have AI but few deliver true, best-in-class data science”. Others bolt on AI applications to existing solutions which are unlikely to have transformative impact like a pure and proven AI solution can. Regardless, the overuse of the term of “AI”, particularly for watered down functionalities, has created confusion and skepticism.
The reality of AI is that some of its most transformative applications for fashion have been around for over a decade. Nextail, for one, has been helping fashion and apparel retailers rethink their core merchandising processes since 2014.
More mature AI systems have had years to learn from real data and continuously improve, making them more accurate and reliable over time.
As Lewis stresses, fashion retail is different. “When a new company claims to have cracked fashion AI overnight, it’s worth asking how.”
Reality # 5: AI isn’t one-size-fits all: Focus and expertise matter
Despite the headlines, AI isn’t a plug-and-play miracle. Its impact depends on two key elements: fit and focus.
Retailers need to choose systems built specifically for fashion and retail to ensure that AI delivers the tangible impact it promises. Industry-specific solutions, unlike generic ones, are trained on the right data and better understand the nuances of fashion that involves seasonality, sizing and the emotional rhythm of collections rather than treating it “like selling batteries or bread”.
Industry focus also simplifies and speeds up implementation as it requires fewer custom developments or modifications to work within a fashion business. Nextail, for instance, impacts in-season processes once inventory has already hit the warehouse without disrupting ERP or other systems. “The rest of the business could carry on blissfully unaware,” as Lewis describes. “It’s not a snake with endless layers.”
Final thoughts: Don’t miss out on practical, immediate benefits
AI for fashion retail isn’t magic. And retailers who hesitate to investigate it based on misconceptions are missing out on practical immediate gains. When applied correctly, AI can deliver on promises and achieve multiple compounding benefits like optimized stock freeing up cash and helping teams hit sales targets faster.
As Lewis sums it up, “AI isn’t magic. Obviously if you’re buying completely terrible stock, no software in the world is going to help you sell it. It doesn’t make the stock look better or smell better, but you know once you’ve bought that stock, there’s no better way of optimizing it”.
___
Does your team need merchandising heroes?
Explore real-life merchandising success stories with brands such as Mountain Warehouse, Guess, and Hackett London.
