What's the Best Way to Approach AI Tech? (Stylus)
“To remain at the cutting-edge, luxury brands must learn to harness AI to pioneer new and meaningful experiences with consumers,” Marc Close, CEO of bespoke apparel platform Bespokify, wrote for the Business of Fashion last month in an op-ed entitled ‘Technology is Eating Fashion.’ With large retailers, including Amazon and ASOS, jumping on the AI bandwagon, Close has a point. But, for those yet to dip their toes into the water, what’s the best way to approach artificial intelligence technology? Here, we look at some companies that are embracing – and creating – AI tech in different ways.
Snaps and Pinterest: Using AI to Empower Brands and Consumers
Chatbots are becoming increasingly interesting to retailers; they’re coming around to the fact that implementing them can be a smart means of cutting costs and offering customers unique, personalised experiences. New York-based company Snaps has helped an array of brands, from Nike to Nordstrom, to bring AI into their businesses with its end-to-end platform for building, activating and optimising mobile-based marketing programs. Snaps’ focus is on mobile messaging, through customer service chatbots (because 9 out of 10 people want to communicate with brands through messaging, according to Snaps’ website), iMessage stickers and personalised emoji keyboards. For those new to the field of AI and conversational commerce, one of the attractive aspects of Snaps’ cloud platform is that it’s designed to make things simple; while brands could employ developers to customise the platform, they’re not needed to create a chatbot, and there’s no coding knowledge required.
At a consumer level, Pinterest is another company that has been banking on AI to overhaul the way users interact with its website – and its been a long time in the making, with it unveiling its first move into visual search technology back in 2015. Now, with its Pinterest Lens feature, users are able to simply open up the Pinterest app and photograph objects they like, in order to browse the site for related topics based on those images. The visual search facility combines machine learning with computer vision to go further than simply showing users similar objects; delving into the broader realm of related topics, too, aids easier discovery of a larger range of relevant items. For example, a user could snap a photo of some organic chocolate to not only find alternatives, but to also discover recipes that the chocolate could be used in. As well as Lens, Pinterest offers other machine learning-powered features on its website, including its Instant Ideas feature, which presents browsers with visual inspiration based on the pin they’re looking at, and its Shop the Look feature, which aids users in finding shoppable products based on fashion-related pins that they love.
Cosabella: Overhauling Digital Marketing With AI
Frustrated by mediocre results they were receiving after working with ad agencies, lingerie retailer Cosabella decided to switch things up in October 2016, and decided to implement artificial intelligence technology as an alternative. The company enlisted Albert, an AI engine developed by Adgorithms, to overhaul its digital marketing strategy. Albert’s machine learning is able to steer Cosabella’s marketing efforts on social media, mobile and email, as well as able to analyse what its competitors are doing – with keywords, for example – thus setting Cosabella ahead of the game. The company’s marketing director, Courtney Connell, explained to Campaign that AI “doesn’t think about people in the same way we [do] … AI will understand subtle manoeuvres that humans would not deem relevant most of the time. And [it] never forgets.” Albert’s ability to spot patterns and trends, and make links that humans naturally wouldn’t, has helped to set Cosabella ahead; it will let the team know when a creative marketing effort is “fatiguing”, for example, so they know when change is needed to keep customers interested.
While adopting Albert has been a bold move (bold by 2016 standards, at least), it’s paid off for Cosabella, with the company reporting a 336% increase in return on ad spend three months in, and a 155% revenue increase in Q4.
Amazon: Using Machine Learning to Predict Trends
Despite changing the way we shop online, Amazon has never been considered particularly revolutionary by fashion fans, who are more likely to do their grocery shop from the ecommerce giant than they are to buy a new dress. However, Amazon appears to have been pouring more effort than usual into its fashion-related output this year. Its Prime Wardrobe clothing subscription service was announced in June. And back in April, the company unveiled Echo Look, a “hands free camera and style assistant” designed to work with Amazon Alexa. As well as giving users the opportunity to analyse their daily getups from 360-degree view, Echo Look also uses machine learning, combined with advice gleamed from expert stylists, to let users know what kind of clothes they look best in.
Last month, at the KDD 2017 conference, Amazon unveiled recent work from its researchers that indicates where the company’s fashion-focused efforts may be moving, and it was heavy on AI technology. Business Insiderreports that one group of researchers has developed an algorithm that learns about fashion trends by looking at a series of images, and can then design its own items in similar styles. Another group, meanwhile, has focused on machine learning that analyses labels attached to fashion images, in order to “deduce whether a particular look can be considered stylish.” This information could, in theory, then be used to provide feedback on outfits, or to provide styling suggestions. Will we really be relying on Amazon machines to help us to stay on trend in the future? Only time will tell.
This post was originally published by Decoded Fashion, who I write weekly blog posts for. You can check out the original version here.
Image credit: Snaps