From Haute Couture to the Bargain Bin: When AI Meets Fashion

From Haute Couture to the Bargain Bin: When AI Meets Fashion

Sales and marketing for all things fashion are becoming smarter every season, with AI now advising shoppers to find the right color and fit for them. Happy customers mean fewer returns, of course, but AI even boosts the shopping experience in brick-and-mortar stores. Fashion designers have for quite some time been experimenting with the latest technology to design and manufacture their wares, and now the clothes themselves get smart. No matter whether you care whether Karl Lagerfeld or a computer, with the aid of machine learning, has devised some fancy garment, one thing is certain: trends will continue popping up, luring customers to seek out the latest looks and hip apparel.

There are two types of shoppers. Those who thrive when trying on as many pieces of clothing as the rack in their fitting room can bear, while others despair over the style and fit of a pair of pants. Casual, slim – or just give up and run off? These questions are even more critical when buying clothes online. AI to the rescue, then! Levi’s, for instance, offers a Virtual Stylist who helps shoppers find answers to their fundamental fashion questions and guides them through the sales process.

The bot uses Facebook Messenger or chat on the Levi’s site to communicate with fashion-minded customers: “Don't know which pair is right for you? With a few simple questions, our Virtual Stylist will find the Levi's you've been looking for.” He or she understands natural language and even finds answers to difficult sizing questions. “TrueFit” enables the bot to offer personalized recommendations for fit and size for every available style. “The Virtual Stylist uses deep apparel data to give you a personalized size recommendation - no trip to the dressing room required.” TrueFit claims it knows what fits a person by comparing data from thousands of big brands around the world with the clothing and shoes that an individual customer likes to wear. Over time, the bot promises to learn which items really fit a person and what type of shoes and apparel they like. [1]

Fashion-Bots made in Silicon Valley

The bot has been built by, a startup based in Palo Alto. CEO  Eitan Sharon explains the vision behind it like this: “The Levi's Virtual Stylist is a tool for customer engagement with the brand, but also for driving sales of products. Our unique visual AI technology is a real game changer, as there are many limitations [that] pure text-based search cannot address.” [2] has also built an AI for Louis Vuitton that helps and aims to inspire customers via Facebook Messenger. Says Michael Burke, CEO of the luxury brand: “At Louis Vuitton, we always think client first. Today, our clients like to be connected to the Louis Vuitton universe wherever they are. They shop in our stores worldwide and are often in contact with their personal shoppers. They also follow us on Facebook, Twitter, Instagram, etc. and find our products on It became evident that we should create the Louis Vuitton Virtual Advisor powered by, a 24h/7 service to be able to fully meet their needs.” [3]

While it’s obvious that you’ll never be able to replace the shopping experience in a Louis Vuitton boutique in Paris, Munich or Rome, AI does have a realistic chance to change the fashion world when it comes to offline shopping, too. More then three-quarters of all fashion retailers plan to invest in AI over the next two years.

Farfetch is one of them. It has built a “store of the future” that automatically recognizes a customer when they come in and is fitted with digital coathangers and smart mirrors. That way, shoppers can switch between any size and color they like, because the mirror will simply show them a green dress in yellow. It’s a nice way to stitch together the online and the offline worlds.

Designed by AI 

Project Muze, a joint fashion experiment between e-tailer Zalando and Google, went even one step further. Its aim was to let machine learning for the first time try its synthetic hand at designing fashion. Users answered questions and offered their creative input on a website to drive virtual designs.
Project Muze is built on top of Google’s open source platform TensorFlow.  At its core sits a “design engine” made up of a neural network. In order to give it some creative smarts, Google trained the network with the color, material and style preferences of more than 600 fashion pros. Over time, the neural network learned which look would resonate with a user. The “design engine” also draws on insights from the Google Fashion Trends Report to come up with even better-fitting suggestions. [4]

The project was, of course, about marketing dressed up as gamification. Users played and therefore engaged with the Zalando brand, while the system suggested products that were similar to the machine-generated designs. With remarkable results. During the first month, the system sketched 40,424 designs, not to mention the splash it made in traditional and social media channels with the hashtag #projectmuze.
It’s giving us a good preview of the future. One day, we’ll be able to bring our designs to life with a 3D printer. And it won’t happen in the far-distant future. Karl Lagerfeld has already tried his hand at it, and back in 2016, actress Kate Hudson presented a printed evening dress by Versace at a gala in New York. [5]

Back to Project Muze. Zalando eventually picked three virtual concepts and turned them into tangible apparel. Models walked the runways at Bread & Butter 2016 in Berlin wearing designs by international fashion bloggers Wana Limar, Anthony Bogdan and Sofia Tsakiridou.

Fashion forward

Here’s what the trend report “The State of Fashion“ predicts for 2018:
“The potential of AI for the fashion industry goes beyond supply chain processes and automation, as it transcends typical machine tasks and blurs the lines between creativity and technology. Leading fashion companies will use it to enhance the creative process, design and product development; they will, for example, use algorithms to sift vast amounts of data to predict which product features customers are most likely to prefer.” [6]

As if to prove that point, the Indian designer duo Falguni and Shane Peacock have used IBM Watson to chart the future of Bollywood fashion by analyzing more than 600,000 runway images of fashion shows and Indian couture.

But not only production and sales & marketing of fashion are getting smarter, the apparel itself is, too. Wearable technology is revolutionizing fashion, particularly the world of sports apparel. Socks to track dementia patients who have become lost, or bracelets that can track a diabetic’s glucose level and automatically inject insulin are fascinating innovations in this realm.

It’s against this backdrop that Deutsche Telekom has launched the “Fashion Fusion” initiative to position itself as an engine of continuous innovation. Designers and startups will present their next round of fresh ideas and concepts waiting to hit the market at the Berlin Summer Fashion Week 2018. [7]



[1] Facebook Messenger: Levi’s Virtual Stylist, (visited 09-01-2018)

[2] Forbes:
The New Levi's Chatbot Aims To Help Shoppers Find The Perfect Pair Of Jeans
04-09-2017, (visited 12-01-2018)

[3] Forbes Magazin:
Louis Vuitton Becomes Latest Luxury Brand To Launch A Chatbot
08-12-2017, (visited 09-01-2018)

[4] Media information Google: Project Muze - Fashion inspired by you, designed by code
02-09-2016, (visited 09-01-2018)

[5] Wirtschaftswoche: Per 3D-Scanner zum perfekten Outfit
05-06-2016, (visited 09-01-2018)

[6] The Business of Fashion: The State of Fashion 2018, (visited 09-01-2018)

[7] Media information Deutschen Telekom: Deutsche Telekom again combines fashion and technology
18-04-2018, (visited 09-05-2018)

Add a comment

your browser is not up to date
to enjoy this website you will need to install a modern browser.
we recommend to update your browser and to install the latest version.

iOS users, please male sure you're running at least iOS 9.

Mozilla Firefox Google Chrome Microsoft Edge Internet Explorer