May 8, 2018
Imagine walking into your favorite clothing store to purchase a new outfit. Rather than combing the shelves, digging through the clearance section, or checking the tags on mannequins, you walk up to a computer that recognizes you and you input what you’re looking for — “outfit for summer.” Using the data from your previous purchases, inventory available at the store, and current fashion trends, an artificial intelligence (AI) algorithm puts together several outfit options for you to choose from. With minimal effort, you find clothing perfectly suited to your taste and needs.
Companies like Alibaba are starting to make scenarios like this one real. For the Chinese holiday Singles Day, Alibaba introduced a new app called FashionAI to several stores. The system scanned clothing taken into a dressing room by a shopper and used the store’s inventory, as well as data collected from fashion experts and designers, to suggest complementary items. Alibaba drove a record $25 billion in sales for that one day.
Marketing and advertising campaigns are assembled more efficiently and more successfully with AI. It helps target the right people. On social media platforms like Pinterest, it’s AI that recommends products to users based on previous “likes.” Other buying and selling platforms are experimenting with AI Computer Vision that allows users to take a picture of something they like, and search for similar products for sale, or take a picture of their couch and find a lamp or table to match the style.
Using AI to boost sales
The amount of money companies have invested in AI technologies has tripled since 2013, and early adopters are seeing better profit margins than their competitors, according to a recent McKinsey study.
As many shoppers move online, they miss the opportunity to see, feel, and try on clothing before buying, and brands miss the opportunity to connect in-person with their customers. But AI applications like FashionAI can bring the two together: they help provide a better online experience, and steer people to the physical store. Because AI can give brands the information needed to understand a shopper’s preferences and fashion sense, clothing stores can better predict what pieces of clothing the buyer might like. In the case of Fashion AI, it combines the ease of online shopping with the tangible in-store experience — making the buying experience digitally immersive and personal.
When customers can’t come to the store, it would be great if they could speak to their own personal shopper. Some can. Many brands are already participating in “conversational commerce,” where customers can use chatbots or their voice-based assistants to communicate with brands in a conversational format from their home — and it relies on AI technology to make it a viable experience. “Brands like Louis Vuitton, Everlane, Burberry, and Nike are using Facebook Messenger,” says Michael Klein, director of industry strategy for retail at Adobe, “which uses artificial intelligence and the ability to take voice to text or use bots to give customers a guided shopping experience.”
This guided shopping experience blends into subscription retailers like Stitch Fix and Trunk Club as well, by combining machine-learning algorithms with human personal stylists to help curate personalized wardrobes for men and women. For example, a Stitch Fix order is processed by 5-10 styling algorithms before it ever reaches a human stylist. Each algorithm has a distinct purpose in the styling process — from matching the stylist to the client based on style preferences to assigning which warehouse assembles the order. Stitch Fix currently employs over 85 data scientists that oversee the entire process.
What about the store’s backend? Relying solely on manual methods of stocking and searching inventory leaves significant room for worker error and upset customers. AI is capable of rapidly searching massive inventories much more quickly, efficiently, and thoroughly than a human can. When companies have a better way to analyze and understand customer data, they have a better idea of which products to stock where.
“Those tags that you have to snip out of your clothing or have to be removed by someone in the store — they’re inventory and loss-prevention tracking devices that are now becoming even more powerful to try to understand and predict supply and demand,” Michael says. AI will help retailers gather and process data from these devices quickly in order to make adjustments to inventory in real-time.
For example, AI could be used to track inventory and predict that a certain product will fly off the shelf in California, but have stagnant sales in New York. With AI/machine learning, the company could accurately forecast stocking more of the item in their West Coast distribution center and much less on the East Coast, improving overall sales, raising margins, and preventing overburdening a store with merchandise that doesn’t move.
Using AI to develop products customers want
Going further outside the retail environment to product development, AI can help analyze consumers’ tastes and the most popular items in a company’s inventory and then convey that information to designers. Designers can create new products they know will appeal to buyers, particularly if AI can help them track trends that may still be in play, or ones that are being exhausted.
For example, if a group of shoppers buys furniture consistently in a certain style, AI can indicate the pattern to product designers, who can create new accessories to match the furniture already purchased.
A group of Amazon researchers has already developed a program that determines whether a piece of clothing is stylish by analyzing images of the latest clothing lines. This feedback is available to designers as they create new clothing designs.
Another Amazon team developed a rudimentary AI fashion designer that generates new designs for pieces of clothing. Using a generative adversarial network (GAN), the program can analyze a wide variety of styles and apply what it learns to design a similar item.
Using AI to improve communication and efficiency
AI can also drive productivity for employees in stores. One new technology, Theatro, allows employees at large retailers to access an AI-powered virtual assistant with their voice to gain information about the store almost instantaneously.
Michael says, “If I work on the floor at Home Depot [in] Canada, I have the ability to connect my Theatro device to an artificially intelligent virtual assistant that talks to me through an earpiece. The assistant can help me understand things like inventory and back-of-the-house operations so that more employees can be on the floor working with customers instead of physically searching for answers in outdated systems in the back of the house.”
With technology like this, employees can work faster and smarter, and can avoid wasting time on menial tasks that can be handled by a computer.
Early AI adopters are also investigating ways to communicate and work globally — like AI-powered translation services that allow a person to speak English into a microphone and have it emerge as Chinese from an earpiece across the world.
A future of enhanced experiences
“It’s all going to become second nature, and we’re not even going to feel it or know it,” Michael says. “The magic of the best technology is receiving a great experience, and not knowing whether it’s through a human being or artificial intelligence.”
The backbone of AI is machine learning with data, and customers are its supply. AI can combine the intuitiveness of human employees with a machine’s ability to analyze massive amounts of data in seconds. As AI is used in more commercial and consumer applications, the possibilities for integrating the technology into commerce and retail experiences will only grow bigger. The benefits of AI are in its ability to understand consumers, improve worker productivity and efficiency, and — maybe most importantly for retailers — boost sales.
Read more articles about how we benefit from the changing landscape of technology in our Human & Machine collection.