trends, fall 2017 fashion trends, delpozo, ibm, watson, artificial intelligence

The past few years of technological advances have led us on a collision course — a merging of couture and computing. The industry has reached a point where machines can leverage enormous amounts of online user data from e-commerce, social media and smartphones to understand and foresee real-life trends, and in return, refresh their products and customize their shopper’s experience.

Yet, only AI assistants have gone mainstream this far in the industry. While chatbots are great, there are many other customer-centric applications of AI in fashion that remain untapped. Here’s how apparel brands can go full-on artificial — and it’s way bigger than Manus x Machina at the Met.

Artificial Craftsmanship

Sure, machines can analyze search trends. But can they actually craft your clothes?

The answer is yes: Amazon’s recently developed algorithm can reportedly “design” clothing by analyzing images and copying the style — giving the e-commerce giant the ability to functionally design its own pieces based off of current trends.

And while Amazon is far from an atelier, with this technology already in the works, it’s easy to imagine that creators could pinpoint a trend, quickly design outfits inspired by it, and then produce and/or ship them on-demand automatically. For this reason, retailers need to think about how they can better predict and react to seasonal trends using AI today — before Amazon completely corners the market.

Similarly, this predictive technology can also provide physical retailers and franchisees with an AI-informed edge when it comes to merchandizing, and stocking for the latest trends — as well as making relevant product recommendations in-store.

The Fitting Camera

While chatbots have quickly gotten popular in the fashion space, helping brands at a variety of price points communicate with shoppers and recommend products based on their expressed needs, they are missing one key thing: Leveraging online user data for “store-like” experiences on mobile.

One example of this is virtual fitting: Gap Inc. recently debuted an augmented reality app in partnership with Google and Avametric aimed at letting shoppers virtually try on clothes through their camera. Customers enter data like height, weight and age in order for the app to calculate their likely body shape and make tailored product recommendations — helping them to have an easier, more body-positive shopping experience.

And this is just the beginning: It’s not a stretch to say that in the near future additional data from social media, chatbot conversations, image searches and more should be incorporated to build a more complete user profile — and use Augmented Reality to give shoppers pitch-perfect recommendations that suit their style and past purchases every time.


Apps and bots aren’t just about virtual try-ons: Plenty of customers value the in-store experience, and they want to look at apparel in person — especially for brands at higher price points.

Department stores in this category have certainly been dealt blows by the digital revolution. But Macy’s is exploring ways to use its higher foot traffic flagships to its advantage, launching an in-store, AI-powered helper called Macy’s On Call, powered by IBM Watson. Customers in the store can chat with the digital assistant, and using natural language processing, Macy’s can understand diverse requests and point shoppers toward the clothing items they actually want to see in the store.

Levi’s new Virtual Stylist chatbot does something similar: It texts with customers — even while they’re shopping — to offer product recommendations based on their preferences. But it can’t be just bots that evolve: Stores themselves have to as well. After all, there’s a reason that click-to-brick pioneers such as Warby Parker and Blue Nile have seen so much success with their digital-first showrooms.

Alibaba’s FashionAI is doing just this in the Chinese market. Essentially, smart screens in fitting rooms can recognize items — a cashmere sweater, for example — through a sensor embedded in the tag and function as a personal shopping assistant right there in the dressing room. But this is about more than matching outfits: This kind of personalized service is only available in physical stores while customers are trying on clothes — and with Alibaba reportedly installing it free of charge at 13 stores in China, it could give shoppers an incentive to visit brick-and-mortar locations when they might not otherwise.

Rebecca Minkoff’s “store of the future” concept is another example of innovation in this category. The label has embraced IoT tech like smart mirrors that can help shoppers put looks together, but the future of this tech has to be able do more than match a top to a skirt.

An important element of this is pattern/body recognition: Retailers should look to invest in tech that recognizes the difference in shoppers’ bodies and make recommendations that suit them. The whole point of AI is that there is no “one size fits all” — and in that respect, it’s exactly like couture.

Omnichannel Performance

Finally, it’s easy to think of AI as a standard product recommendation engine. But it’s also serving to connect the dots between online and off-line in retail — from footfall and online traffic to transactions.

Fashion brands can now use social apps such as Facebook, Instagram, or even Snapchat and Whatsapp, to target users contextually and situationally, as well as to attribute in-platform marketing to actual sales, whether it’s in-store or online. Similarly, AI-powered CRM systems can now understand the entire purchasing journey, regardless of digital and retail touchpoints, helping brands push their message where and when it’s most relevant.

After all, when it comes to fashion, consumers don’t distinguish between online and offline as long as it fulfills their needs.

As a global executive for WPP agency Kinetic, Benjamin Lord serves as a strategic marketing advisor to major global brands including Calvin Klein, Chanel, LVMH and more. He can be reached at: