How Brands and Startups Are Using AI to Help You Get Dressed

Chatting with a stylist (some human, some not) is the hot new thing as retailers develop their "conversational strategies." What's next?
Author:
Publish date:
Social count:
493
Street style at London Fashion Week. Photo: Edward Berthelot/Getty Images

Street style at London Fashion Week. Photo: Edward Berthelot/Getty Images

If you're reading this article, chances are that you're pretty aware of the fact that many things are moving from the physical world to the digital one — from editorial content to retail. But sometimes, those online experiences still leave something to be desired; a trend we've noticed over the past couple of years is retailers harnessing technology in an attempt to mimic the level of customer service and personalization you might get from a really good, attentive salesperson IRL.

While shopping online is supposed to be convenient, it can often be overwhelming. Many online retailers boast tens or hundreds of thousands of brands and SKUs, and if you don’t know exactly what you're looking for, and how it will fit you, the experience can be pretty frustrating. To mitigate that, retailers with the resources to do so are working to use data collection and, in some cases, AI to create more personalized shopping experiences — i.e., showing you products it thinks you will like based on what you've purchased before, sort of like fashion's version of Spotify Discover Weekly or Apple Music's For You tab.

But some are taking this a step further in what is a sort of next-level convergence of two fashion-tech phenomena we wrote about in 2014 and 2015, respectively (technology moves fast!): personal styling startups and retailers' use of mobile messaging for customer service. While many of us feel confident in our personal style, others find getting dressed in the morning to be a daunting task (and we all have those days, don't we?). Today, retailers and tech companies want to help with that by styling you via a chat platform — and not all of them are even trying to sell you clothes.

Many of these "virtual stylists" are, in fact, chatbots — something many retail brands have experimented with over the past couple of years to mixed results. These live primarily on Facebook Messenger given its over 1.3 billion users and the tools it offers developers to build their own interfaces. 

In September, Levi's partnered with Mode.ai — a Silicon Valley-based startup that builds chatbots for retailers — on Levi's Virtual Stylist, which aims to help users find the perfect-fitting pair of jeans. It lives both on Messenger and the Levi's website, and will ask you your preferences when it comes to fit, rise, amount of stretch and wash; it also asks you what size you are in another brand (say, Madewell) to determine your size in Levi's and suggest the right pair. "To figure out how to even navigate through a website is difficult for many customers initially, and [that's] not to mention the fact that you can't easily say, naturally, your preferences for jeans," explained Mode.ai SVP of Business Development Karen Ouk of the problem they were charged with solving. "They found it very difficult to address the experience [an in-store associate] would provide to their customers." Given the unique challenges to finding denim, they wanted a natural, conversational way to help shoppers.

Mode.ai also operates its own Messenger bot-styling service, asking questions to determine what you're looking for and then directing you to items it thinks you'll like, using your feedback to find more personalized options as you go.

Related Articles

How Artificial Intelligence Can Help You Build Your Own Personalized E-Boutique

How Western Fashion Brands Are Using Social Media in China

How Retailers Are Using Mobile Messaging to Change the Way We Shop

A similar and, according to Facebook, very popular chatbot is Epytom, a personal styling service not tied to any other brand that encourages users to shop their own closets rather then driving them to a point of sale. It helps style them using clothes they already own based on their preferences, as well as elements like weather and what they have to do that day. Through surveys, founder Anastasia Sartan has gathered some insight into how useful personal styling bots can actually be. 

"We learned that people crave the simplicity of having an outfit already picked out for them when they wake up. Except for true fashionistas, most of us are confused and frustrated with the overwhelming fashion choice available to us," she told me over email. "Sixty-four percent of our users wear what we recommend them more that twice a week, 20 percent three-to-five times a week, and all of them tell us we make their mornings easier and less stressful." She's also focused on not annoying users with notifications. "Eighty percent of our users love the bot format; they are glad to regularly receive notifications from us because we give them practical, applicable advice," she says. "Without personalization and direct value, bots become just as annoying and spammy as unsolicited emails. This is not a long-term strategy and it doesn't drive the chatbot industry forward."

While Eptytom seems to have largely succeeded in engaging its 500,000 users, a common problem with bot services is disappointment with, well, the fact that it doesn't always feel like chatting with a human. "While we are constantly getting more precise in sending you an outfit with a piece you want for your agenda and weather and that you'll like, we still can't respond to the questions like, 'What do you think of wearing sandals with socks?' and it disappoints some users."

By the year 2020, over 80 percent of retailers are predicted to be using chatbots in some capacity, and though many believe that chatbots represent the future of retail, there's been an ebb and flow in their success rate. While several digital-first retailers were quick to launch chatbots early on, some have abandoned the concept recently, finding that they didn't meet the level of personalized interaction and assistance users expected — including Everlane and Spring. Trunk Club, a subscription styling service, posted an article last year about how it prototyped a styling/shopping bot and then decided to scrap it. The consensus among experts seems to be that they need to truly add value and utility for people to opt in, and perhaps, some of the aforementioned styling services do.

Not all of today's virtual stylists are bots, however. Some connect users with real humans.

Nordstrom has been an early adopter in this space, launching TextStyle in 2015, through which users could chat with and purchase from sales associates or "personal stylists" via a messaging platform called Next, which doesn't seem to have taken off. Its latest experiment in this arena, which launched alongside the retailer's new Nordstrom Local inventory-free concept, is Style Boards. It allows salespeople to create and send customers Pinterest-like boards with personalized fashion recommendations, perhaps for a special event or trip. Customers can converse and purchase items directly through the app.

But the next generation of these services may not require any typing at all. This year, Amazon launched Echo Look, an Alexa-enabled device that also has a camera. It has a feature (or "skill," as Amazon calls them) called Style Check, which you can ask, verbally, for weather-based outfit suggestions. And just this week, social commerce platform Poshmark announced (alongside its $87.5 million funding round) the launch of its own Amazon Alexa skill: Stylist Match uses AI to connect shoppers with a few of Poshmark's millions of Seller Stylists who, when asked ("Alexa, ask Poshmark to style me."), create customized, shoppable looks culled from their inventory for a variety of occasions, from "Date Night" to "Casual Weekend." It uses data to specifically pair shoppers with sellers that reflect their style preferences.

"I don't think there's a true alternative to what we are creating except if you are shopping in the thousands of dollars at a Neiman [Marcus] or places where they can afford to apply a physical stylist," Poshmark Founder & CEO Manish Chandra tells me over the phone. "Otherwise, everything else is, I would call it personalized, but not personal." Both of these skills use a combination of AI and real, human stylists. (Though, it's worth pointing out that in the startup world, the term "stylist" is defined a bit more loosely than it is in the fashion world.) 

Rachel Arthur, a journalist and consultant focused on the intersection of fashion and technology, feels this combination will be the most successful. "At this point in time, that's the sweet spot — offering something as broad as possible by integrating machine learning and then enhancing that experience by ensuring there's a real person on the other end actually delivering the final piece of it," she says.

There are benefits to both sides. With bots, it's scalability. As a retailer or styling service brings in more customers and users, it can meet their demands without having to hire more people. "Bots provide an opportunity to make expert knowledge and service highly scalable and accessible to a much wider audience without the financial and infrastructure constraints," says Sartan, noting: "Of course, different business needs and models require different approaches and we believe that there is room in the chatbot space for human-human interaction, either for an even more personalized concierge level of service, or for solving a specific issue in real time."

Chandra feels that people want an experience with real people. "Where we're using the AI and deep learning is in the match, but the actual styling is still done by people and I think that's what people want," he says. "A lot of these other technologies are meant to, in some ways, eliminate people so that you don't have the people thing, but I think it seems a little dystopian to me right now."

Both, if done successfully, drive personal relationships and customer loyalty, which is why this is surely just the beginning of brands developing what Jonathan Shriftman — Director at Snaps, a startup similar to Mode.ai that develops bots for companies like Nike, Coach and Macy's — calls a "conversational strategy," whether it's through texting or voice-enabled assistants; humans or bots. 

"With the proliferation of Alexa and Google Assistant, we're going to be talking to a lot of our hardware, too. We're going to a lot of brands and saying, 'You guys have an Instagram strategy; you guys have a Facebook strategy; you guys have a Youtube strategy, but what's your conversational strategy? Because this is the next big thing that's happening, and you need to be where your customers are,'" he says.

Chandra sees future iterations of Stylist Match where you can ask more detailed questions and execute a full styling conversation while driving to work, either via a future mobile version of Echo or, say, a Tesla with its own voice-enabled assistant. Ouk points out that the next Oculus virtual reality headset will be much more affordable and likely inspire many more brands to develop shopping and styling interfaces for VR.

But let's not get ahead of ourselves. With the exception of a few early adopters like those mentioned here and others like Burberry and Tommy Hilfiger, fashion is famously reluctant to embrace new technology. "There are few retailers that are really true innovators... who are willing to test out these technologies ahead of the curve and really innovate and push the system, but many, many more in fashion and retailer are followers," says Ouk, who has also worked at Google. "They'll see the innovators have done something and many times wait for the results to see what's going on and make sure they are positive results before they make any changes."

That said, given the dire state of brick-and-mortar retail right now due to evolving shopping habits, brands are abnormally proactive about identifying new ways to drive sales. Says Ouk: "I think those kinds of forces are pushing them to maybe innovate, even if it's not part of their DNA."

Want the latest fashion industry news first? Sign up for our daily newsletter.