Fashionably AI

This summer’s wedding season required me to buy a new suit. I vowed to be adventurous and buy a color I normally never would have considered. Alas, I opted for a little more movie-theater usher and a little less Jidenna. Had I known about it at the time, I probably would have used Eison Triple Thread, a company that specializes in creating made-to-order suits.

Working with someone to create a suit can be a hard enough task. You have to consider the occasion the suit is for, body type, taste and other relevant factors. And what other suit company or department store doesn’t already do that? To differentiate itself from the crowd, Eison Triple Thread launched FITS, a web application that creates tailored looks based on clients’ lifestyles and musical preferences.

Eison founder and CEO Julian Eison was the fly kid on the playground and says his parents instilled in him a sense of presentation and to be his best when he was out and about.

“In terms of style and color I was super deliberate about what I wore,” he says. “I was the kid who collected Jordans and wanted to be fashionable because I just cared. I think through that process, and as I grew, I just started to embrace it.”

After six years in private equity, where he says he was able to see tech’s flow from the buy side and the sell side, Eison decided to combine his love of fashion and interest in tech. In 2014, he launched Eison Triple Thread from the garage of his San Francisco home to try his hand at creating an alternative to suit-buying at conventional big-box department stores.

“When we first launched the business, it was about visualization,” Eison says. “How can you visualize your body and think about something going on your body that fits you well?”

But Eison Triple Thread isn’t the only suit company that wants to outfit its customers in sleek styles in a made-to-order fashion. The likes of Indochino, Bonobos and Stitch Fix, all of which came before Eison Triple Thread, ultimately have the same goal. So what’s a suit company do to strike a difference between its competitors? Why, integrate artificial intelligence and Spotify data, naturally.

“Music is at the core of a lot of everyday life; it knows no boundaries or color, and it reveals something about us that we may not know that we kind of project onto people,” Eison says. “So we’re trying to get to the core, the unadulterated piece, and that’s music, and it drives a lot of our decisions, selections, identities and moods.”

During the onboarding process, users first log in to the FITS system with their Spotify credentials and take a lifestyle quiz. Questions include in which industry you work, how you dress for work, what your work commute is, how you spend your free time and which word best describes you. Eison says they can start generating data from this basic information.

“We’ve turned that into a lifestyle quiz that aims to reveal as much about a person in terms of their fashion, their interests, their preferences and how they typically like things to fit. That goes into our analysis and allows us to home in on this fit and this style.”

While you’re busy thinking about yourself to the best of your ability, FITS is trolling Spotify through its API to gather data about your musical tastes: genre, when you tend to listen to music and for how long. The process from beginning to end takes only about 15 minutes — unless, like me, you have a hard time selecting just one word from a list of four to describe yourself. Reflective, intense, upbeat, energetic: I am all of these things.

Once you complete the quiz, the web app returns a list of “looks,” as Eison calls them, based on data gleaned from your best answers to these questions. The looks come from a collection of images that Eison and product director Dario Smith curate regularly from the internet based on styles they deem worthy. Eison tells me they currently have 3,000 images in their database and curate additional ones seasonally to kick back to customers on a regular basis.

They pull the metadata of photos, including color pairing, assumed cloth texture and other similar data, which the algorithm uses, Eison says. In the next release, he said the company will be able to identify skin tone for those who upload the required photos. In addition, the company uses available photo metadata to understand geography of fashion. When available, Eison says, they are able to gain insight into local fashion and trends to further tune the algorithm.

“If there are X amount of styles, we want to make sure we have representation,” Eison says. “We can aggregate all these images and then serve those periodically based on how important or relevant they are.”

For my part, I answered the questions while Spotify worked in the background to make sense of my musical predilections: showtunes (your Hamiltons, your Ragtimes, your Cabarets), Jidenna, Calle 13, selections from Moana (yeah, that’s right), Nathaniel Rateliff & the Night Sweats and a smattering of old R&B.

The result was a list of 25 photos of men of varying ages, races and sizes in a wide range of suits pulled from the Eison database (see five of them below). I was excited about most of them, although there were a few too many double-breasted ones for my liking. That’s on me, I suppose, but I don’t think that’s a look I can pull off. Or maybe that’s the point of a system like this: To present something to someone that he or she might not think they’d ever look good in or visualize themselves even wearing.

Once you select the look you want, there might be further details to tend to, such as number and style of jacket buttons, button-hole color, the color and fabric of the jacket lining, waistband style on the pants and anything else you can possibly think of. One thing I could see in the future is the ability to place these looks on a picture of myself.

Once you make all of these very permanent decisions, you then have to be measured. Or measure yourself if you opted to do this at home. I was in the Eison studio, so Smith did the honors, measuring me in places I never thought needed to be measured. For instance, they noted posture, as well as the way my arms rest on the side of my body. Suddenly I realized why the clothes I’ve worn my entire adult life never fit me very well.

About two weeks later, you have a suit that you picked out not from a rack but one suggested for you based on your lifestyle and musical tastes. And it will fit only you. My suit fits. But because it’s tailored with my measurements, I’m not so surprised by that. The treat here is the unique application of Spotify and machine learning. Having the FITS system tell me to avoid buying a light gray suit is the permission I needed to step outside of my fashion comfort zone and don a look I most likely never would have otherwise. 

Not bad for a music-streaming platform and a little AI-style effort.