How AI Is Reshaping Fashion

Marketing

Fashion brands are using AI to create more targeted marketing opportunities. For example, a retailer selling high-end men’s outerwear uses AI to identify which products (such as coats and scarves) present the best cross-selling opportunities and to understand that customers in Chicago, Illinois, are more likely to buy coats with longer waists for battling severe wind and cold in the winter, whereas customers in Phoenix, Arizona, probably are seeking a more short-waisted coat because they’re likely interested in outerwear that doesn’t get in the way of outdoor recreational activities, like skiing. In other words, people in Chicago are trying to fight the cold, whereas people in Phoenix are probably trying to play in it while on vacation. This retailer uses traditional AI to identify such less-obvious insights, and it uses GenAI to create customized marketing emails with personalized subject lines as well as marketing copy adjusted for each potential consumer.

One luxury label used GenAI to generate photorealistic backgrounds for fashion shoots, avoiding the enormous costs of going on location. Another label used GenAI at its fashion show booth to provide on-brand responses to style-related questions from attendees.

A US-based upscale shoe brand is using GenAI to rapidly scale out hundreds of A/B tests on its website, often making only very small changes, which allowed it to discover that product descriptions could be better placed to complement accompanying visuals.

Ecommerce, Sales, and the Customer Experience

Fashion retailers are using AI to vastly improve customer service. For example, one clothing retailer uses the technology to aggregate purchase data and other information about customers as they call or log in to its commerce site, and then it uses GenAI to provide scripts that call center agents can use for help with returns or for suggesting assortments. The retailer’s use of AI in this manner helps improve customers’ entire experience, from when they first consider making a purchase (using GenAI to make compelling assortment suggestions) to ensuring any returns or exchanges are handled as smoothly as possible. AI-based analytics can also provide guidance about why a particular customer likes a given brand, presenting only options most likely to end up in a sale.

Fashion brands also are beginning to use GenAI to generate product collections based on selfies that customers have uploaded to their commerce sites, helping customers make buying decisions and ultimately reducing returns. One high-end apparel retailer uses GenAI to create more compelling in-store signage and product displays, boosting sell-throughs.

Another online fashion retailer has begun using a GenAI-powered chatbot to interact with customers online, respond to questions, and act as a personal shopper.

Supply Chain and Logistics

Retailers are using AI to improve how they manage the movement of goods as they arrive at distribution centers and individual stores. Variables include the presentation requirements and the number and provenance of different goods as well as dates and times that merchandise can be expected to arrive in the back of the store. AI can help schedule support staffers who bring goods from warehouses to store rooms and from storage to the sales floor. AI also can help ensure that the right fixtures (such as shelves and T-stands) are in place based on new arrivals at any given moment. Fashion brands also are using AI to identify the most cost-effective distribution centers based on store inventory allocation decisions (which also are aided by AI).

Some fast fashion brands are using GenAI to create what-if scenarios for alternate supply chain routes in case of a natural disaster or other disruptions. For example, one label is using AI to help identify alternate routes to offset disruptions to shipping lanes around the Suez Canal.

The fashion industry is also turning to AI to help reduce fabric waste. This is a significant challenge for a cut-and-sew shop because patterns can’t overlap as sleeves, the shoulder, and cuffs, for instance, are sewn together (ahem) seamlessly. “Say I’ve got X number of yards of some pattern, and I’ve got an order for, say, 10,000 items. And the pattern gets laid out in the most efficient way, where there’s no crossover to the pattern, yet I’m able to cut every sleeve, placket, shoulder, and cuff that I need, all with absolute minimal waste,” says Flinn, recalling his time at Neiman Marcus. “That’s some in-depth math that was way over my head. That’s what we need the AI for.”

Merchandising and Inventory Allocation

AI can help buyers and merchandisers allocate a mix of fashion items to hundreds or thousands of different stores, taking into account the local demographic mix and demand history, the dimensions and layout of each store, shelving fixture requirements, storage availability, and available inventory. “There’s some very interesting utility around AI and visual in-store presentation,” Flinn says. “For example, I know coming in I’ve got a certain number of sweaters that have to be folded, pants that have to be hung, and dresses that have to be hung—and the dresses are long, so they need taller fixtures.”

Based on this data, AI can recommend changes to assortments and the purchase of new store fixtures, and, knowing how each store is laid out, recommend changes to the visual presentation for each one. And if some merchandise hasn’t made it into the store yet, AI can make recommendations to fill out store shelves with older inventory that makes the store look more presentable to shoppers and that has strong affinities to what’s already on the shelves. “AI in planning is really critical because it takes the science out of the users’ hands and allows the users to focus on the art that exists in fashion,” Flinn says. “There’s still an art to identifying the right print, the right weight of fabric, the kind of button you want—all of the details that the science can direct but not ultimately decide upon.”

One specialty retail chain is taking the output of thousands of Monte Carlo simulations to better understand why some stores may have underperformed in the past, helping them avoid unnecessary out-of-stocks and improving sell-through rates. Executives and managers still use their experience and eye for detail to identify attractive merchandise, but they’re freed from the tedious aspects of determining how many of an item to buy and where to allocate it. “The user doesn’t get caught up in the science,” Flinn says. “The user takes the guidance and the guardrails provided by the system and goes into market and finds the 15 dresses or whatever that the plan says to buy.” But it’s still up to the user to identify the right merchandise.

AI also can help fashion retailers identify stores, based on location and demographics, that have the potential to sell higher volumes than they have in the past. “Maybe Beverly Hills has always been my No. 1 store. Well, it’s a bit of a self-fulfilling prophesy,” Flinn says. “If you send that store most of the inventory, it’s always going to be the No. 1 store. But imagine what you could be doing if you started to move inventory elsewhere. That’s where AI can help.”

In-Store Operations

Retailers of high-end and fast fashion typically operate with very narrow margins. AI is emerging as an effective means of improving those margins by making store operations run more efficiently.

For example, one retailer uses AI to correlate not just high performing salespeople with days of the week and times when store traffic is highest, but which sets of employees work together best, to make sure high performing teams are on the store floor during high-traffic times.

Another brand is using GenAI to help its merchants understand why sales of winter sporting goods are down despite cold temperatures and an abundance of snow. It’s also using AI to help decide which new products to add to its homepages and which categories to put on sale—and using GenAI to automatically make the necessary changes to published prices and marketing copy with a simple prompt (“i.e., discount all my ski boots and accessories by 15%”).

Product Development and Innovation

Fashion retailers rely heavily on creativity to keep customers returning. Nowhere is the “art” of fashion more highly prized. But even here, the industry is turning to AI for deeper insights into trends that can influence design and product development and to create mood boards and other design templates in mere seconds that would ordinarily take hours or days to accomplish.

For example, a US-based retailer of sports apparel used AI to determine that a particular line of skirts didn’t sell as well as others because of its type of zipper. That led the retailer to ask the designer to switch to more stylish (and more expensive) zippers. “AI can pull out the commonalities that aren’t easily seen,” Flinn says. “And then what design should do is take that knowledge and do something with it.”

A European fashion label is asking its GenAI model to iterate new design ideas based on images of collections from its previous season, results from which its designers refined using voice prompts.

One US men’s luxury brand uses GenAI within a virtual reality game on its commerce site to help consumers refine their color palettes and create more realistic fabric textures and shadows. Consumers then can purchase these designs, worn by players’ avatars, in-game. Such VR games are expected to pump an additional $19 billion in revenue into the luxury fashion market in 2030, according to Morgan Stanley, as well as help brands tap into a younger generation of consumers and introduce them to the largely male cohort of gamers.