AI is playing an active role in blurring the boundaries between the in-store and ecommerce

AI will increase convenience by increasing efficiency across the entire retail ecosystem – that means both online and offline. (Photo by – / Charlotte McCurdy / AFP)

AI is blurring the lines between in-store retail and ecommerce

Putting artificial intelligence (AI) to work within the fashion trade is nothing new really. And technology has frequently been pressed into service to improve retail experiences, with new innovations supporting brick-and-mortar outlets as they pursued higher profit margins. And all along, ecommerce platforms have improved their services with the help of digital tools.

Prior to this, customers might have been able to order custom-fitted clothing from physical stores, 3D printed to their body sizes. In its first virtual show, the Shanghai Fashion Week tapped the potential of live-streaming, 5G technology, and the digital marketplace to bring the show to a larger audience than ever before. Even the apparel supply chain has become more transparent with the use of blockchain tech.

When it comes to fashion, AI has been successfully empowering both online and offline channels. During China’s Singles Day weekend, for example, Alibaba’s FashionAI deep learning kiosk gives suggestions on what else customers should buy from the retail outlet. And AI-powered augmented reality applications allow online shoppers to virtually ‘try on’ clothing from an ecommerce portal.

When did AI get so popular in fashion? “The technology is starting to become good, and easy to use for both the fashion retailers and the end-users,” says Gijs Verheijke, the founder and CEO of highend streetwear e-store OX Street. “In fact, people already use more AI than they think.”

But in light of recent world events, the lines that traditionally defined what is in-store shopping and what is online shopping have become fuzzy. And AI is playing an active role in blurring the boundaries between the physical and the digital. “It’s most easy to imagine recommended products when you think of AI, but I think the real revolution is happening when the products coming out are entirely built around AI. What does the TikTok of fashion look like?” wonders Verheijke.

“There is so much opportunity now to marry media, social, gaming, and shopping. The whole categorization of companies between retail, marketplace, and media is going to get upended, and a lot of the parts of that disruption machine are powered by AI.”

And as with distributed blockchain tech, AI will increase convenience by increasing efficiency across the entire value chain, continues Verheijke. “You can think of sourcing products, design, warehousing and logistics, quality control, authentication and inventory management.”

AI also has its role to play in the biggest unsolved problem in fashion ecommerce – product returns because of sizing issues. It’s estimated that between 30% and 50% of all purchases are returned. This cuts heavily into the retailer’s margins.

“AI, in combination with Augmented Reality certainly has a key role to play in solving this problem,” asserts Verheijke. “Lowering returns will indeed lower costs and environmental impact. It is also possible that better predictions about what people will buy could lead to lower waste.”

Sustainability advantages aside, AI is by no means perfect, relying on the input to form decisions which just like humans, might not always be the right ones. “There are certain issues that come from having AI and traditional computing work side-by-side, similar to the issues that come up around mixing self-driving and human-driven cars on the same roads.

“AI is much, much better under the right conditions, but lacks the capability for judgement,” Verheijke says. “It relies entirely on what you feed it, and like with any model, garbage in = garbage out.”

One of the major plusses of AI in fashion ecommerce is personalization – the more data it has on you, the better it can search and recommend garments suited to your tastes. But as good as it can be, over-optimizing can be a challenge. “I have enjoyed various kinds of ‘crate-digging’ activities in my life: Discovering underrated sneakers, new and under-rated clothing brands, or great lesser known music,” muses Verheijke. “I have yet to find an AI powered discovery experience that successfully facilitates crate-digging and keeps the discovery experience fresh and enjoyable.”

Like browsing the aisles in a brick-and-mortar store, the discovery experience is vital for helping an online store keep its audience engaged and coming back to purchase more. Are there other parts of the fashion shopping experience that AI can further revolutionize? “There are two that come to mind: the first is the breakdown of relational databases, and the other is design,” notes Ox Street’s Verheijke.

How an AI product database works for an ecommerce platform like Ox Street is through a relational database where product attributes like an SKU, brand, color, high or low cut, are recorded. “Those attributes are still human-defined. What AI is uniquely able to do, is find attributes that are a lot more important, but that doesn’t have to make any sense to a human,” he elaborates. “So instead of setting up the database relations yourself, you dump all your data into one big ‘data lake’ and let the AI take it from there.”

When it comes to design, Verheijke predicts that affordable, ‘fast’ fashion houses will begin to replace their designers with AI. “They rely on identifying new fashion trends quickly, copying the designs almost but not quite, and bringing them to market super quickly. AI will be able to excel at those tasks soon,” nods Verheijke.

But the real surprise comes later, he says, when AI ceases to merely copy designs and instead has assimilated enough information to step into actual couture. “Later, we will also see AI entering the ‘real’ fashion design space, and becoming a real force in determining the direction of designs.”