How Levi’s is using AI to change its jeans business

10 Nov 2021

Dr Katia Walsh. Image: Levi Strauss & Co

Jenny Darmody spoke to the chief strategy and artificial intelligence officer at the popular jeans company to find out how it uses AI.

AI has become a game changer in many industries. For example, Qatari researcher Abdulaziz Al-Homaid is using AI to discover more about population-specific risk factors for diseases such as diabetes.

In marine science, researchers are using a specialist microscope that leverages the power of AI through an attached iPod Touch to detect harmful algae. And in the financial world, AI is helping to improve fraud detection.

But what about the world of fashion? How would a well-known clothing company leverage AI and data analytics?

To find out, I spoke to Dr Katia Walsh, chief strategy and artificial intelligence officer at Levi Strauss & Co, the company best known for Levi’s jeans.

Walsh has more than 20 years of cross-industry experience combining strategy, data, analytics and machine learning, having been the first chief global data analytics officer of Vodafone Group, along with holding several other data analytics leadership positions in various companies.

I caught up with Walsh during Web Summit 2021 to find out what the role of AI is in a clothing company such as Levi Strauss & Co.

“Levi’s has always been a company of innovation, dating back to its beginning 168 years ago,” she said. “It’s now a company of AI because artificial intelligence is a capability that is part of today’s modern business.”

Walsh’s own role is a fusion of strategy, digital capabilities, data, analytics and artificial intelligence. She said artificial intelligence is embedded in three key areas within Levi Strauss & Co: customer experience, commerce and product creation.

‘We now have van Gogh’s Starry Night trucker jacket’
– DR KATIA WALSH

From a customer experience point of view, this tech helps the company personalise your shopping experience based on previous browsing history. Another feature, which has just rolled out in the US and is expected to come to Europe in the coming months, is visual search.

“You can upload a photo of something you like, it could be from a television show that you’ve seen, it could be something that you drew because you want it, and we will actually match that to show you what we have that’s closest to that,” said Walsh.

Automating commerce

When it comes to commerce, she echoed a sentiment common among those who work in AI and data analytics: everything that can be automated should be automated. And so robotic process automation is used at the company to automate mundane tasks, freeing employees up to concentrate on other work.

Another area within the company’s commerce strategy is optimisation, which involves a complicated recommendation engine designed by Levi’s during the pandemic when stores were shut. This engine uses multiple data points to optimise distribution to get orders to the customer in a quick, cost-effective way.

While Walsh said it is usually cheaper to ship an order from a distribution centre than from a store, this sophisticated engine is able to weigh up whether or not an in-store item may have to be discounted down the line in order to be sold. Therefore, it may be better to ship it from a store now, rather than from the usually cheaper option of a distribution centre.

Walsh said the company also uses AI to predict the demand for its items.

“We have created a livestreaming repository of data incorporating all kinds of data, such as transactions, browsing behaviour of consumers but also weather, climate, economic outlooks, epidemiological models, social media trends, fashion trends, competitive intelligence. All of this is streaming in our repository and then we apply machine learning to predict the demand for each individual item that we design.”

AI-powered design

The third piece of the AI puzzle at Levi’s is product creation, where Walsh said AI-powered design now factors into the company’s clothing.

“We have a designer who is joining my team who has actually created a style transfer algorithm. This is a powerful neural network that he created. He’s feeding thousands and thousands of art images into that neural network.”

These images can then be superimposed onto Levi’s products. “We now have van Gogh’s Starry Night trucker jacket,” said Walsh. “We haven’t made them yet, we have the designs, but we plan to manufacture those and that’s an example of smart creation powered by AI.”

A design of a Levi's trucker jacket with van Gogh's Starry Night imprinted on it. A faded image of Starry Night is in the background.

A design of Levi’s trucker jacket with Vincent van Gogh’s Starry Night imprinted on it. Image: Levi Strauss & Co

How data has changed the fashion industry

One of the biggest changes Walsh has seen throughout her time working with data analytics and AI is that this technology is now important and cool.

“It’s almost like revenge of the nerds. I’ve been a nerd all my life, I never envisioned myself working in a fashion company. But now it powers every company, now it’s the essence of modern business, this combination of digital data, artificial intelligence and analytics.”

Other changes Walsh is happy to see in recent years include the attention to privacy, advancements in the area of responsible AI and the ability to disrupt the fashion industry through technology.

“It used to be that a designer would have a great idea and they would make it happen and they would manufacture certain things and then you have to figure out how to sell it. Today this paradigm can be entirely flipped. We are actually on the verge of mass customisation at scale where you no longer sell what you manufacture, you actually flip it around and manufacture what you sell.”

Walsh added that she finds this particularly exciting because it could make the industry more sustainable and less wasteful.

“You manufacture what you know consumers want, will buy and wear and they have already about bought it, so there’s no waste any more.”

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

Jenny Darmody is the deputy editor of Silicon Republic

editorial@siliconrepublic.com