Levi’s A.I. bootcamp is helping designers create denim of the future

“I remembered growing up in Bulgaria where Levi’s was so much more than a garment; it was a flag for freedom,” Walsh said. “I remembered this picture of young men and women sitting on the top of the crumbling Berlin Wall, all wearing Levi’s. To bring this expertise to a company that is such a platform for doing good in society, it can help an entire industry.”

Levi’s has long held a history of innovation. It was the original startup of San Francisco, where the company began designing denim jeans in the 1870s. Today, though, the brand is far from the manual cutting process of the 19th century; now, A.I. is being used to create fashion designs, optimize production and manufacturing processes, influence financial decisions, recommend labor and staffing decisions, guide the go-to-market process, and perhaps more importantly, empower team members from the back rooms of the retail stores to the Fortune 500 C-suite.

“Every company today is an A.I. company, whether it realizes it or not,” said Walsh, who serves as a senior vice president and chief strategy and artificial intelligence officer. “That’s the reality. We experience A.I. in our daily lives, and it’s so critical for any business to integrate it into its daily operations.”

This is the first time a role like Walsh’s has existed at Levi’s and according to her, the first in the fashion industry as well. Part of her task is to instill A.I.-driven technology and methodologies throughout the company. Some aspects are obvious: machine-learning guides the lasers that finish a pair of jeans with a particular pattern and offers personalized product recommendations to e-commerce consumers based on their previous browsing behavior.

During the pandemic, her A.I. work became extremely important to the health of the overall company. At a time when fellow denim companies began discounting products to move merchandise from closed brick-and-mortar stores, Walsh and her team applied machine learning to historical data to predict at what price each garment would sell at what point in time in what specific part of the world.

The Levi’s team began shipping orders from local stores—lowering shipping costs and getting to the consumer’s hands faster—rather than centralized distribution centers. And they used that data plus social media, epidemiological models, and economic outlooks to predict consumer demand for products by quarter for next two years, something she knows will optimize production and reduce waste. Walsh said that they currently have one of the best financial margins they’ve had in years.

Where her role has taken on a truly innovative stance is through her internal A.I. bootcamp, where her team has trained more than 100 employees—with no background in tech or coding—in the basics of artificial intelligence and machine learning over the course of eight weeks. Some of these employees have worked for the company for 30 years. Nearly two-thirds of the recent class were women. One was a stylist in a premium outlet store. Another restocked jeans on the shelves of a retail shop. All together it’s creating opportunities for team growth as well as imaginative problem-solving.

“The beauty of taking people who know the company and industry is that they have been facing these problems for years; they just didn’t have the tools to solve them,” Walsh said.

She cited a recent example from the Levi’s distribution center in Las Vegas. At least once a day, some kind of equipment would malfunction or fail entirely, leading to downtime in fulfilling orders. The bootcamp team worked with Walsh to create a model that would predict when equipment would fail. They then designed an app that automatically dispatches technicians for preventative maintenance.

Now, the team isn’t losing precious time waiting on equipment. They are functioning at full capacity.

Similarly, on the creative side, Ron Pritipaul, an associate data project manager for computer vision, took his learnings from the bootcamp to the thread color matching process. The manual task is as tedious as it sounds. In fact, so laborious that it would take Pritipaul some 20 minutes for a single thread under specialized equipment with precise lighting in the office. Now, thanks to an algorithm he created using color spectrometry data, he can employ a tool that matches threads at 85% accuracy. The manual undertaking would take the team up to two days each season. Now it’s done in about 20 seconds.

But that’s not all. He applied his bootcamp education beyond problem-solving to forward-thinking permutations too. At Levi’s, he originally worked on the men’s fashion design team before broadening his role to one in technology. A proclaimed lover of denim, Prtitpaul wanted to explore the future of denim clothing conception. For that, he adapted a style transfer algorithm to develop an ongoing experiment making A.I.-inspired garments. He is currently working on a trucker jacket inspired by Vincent Van Gogh’s The Starry Night, which features a specific, hand-painted design with fine-tuned control. What was once reserved for couture fashion ateliers may be done at a mass scale sooner than we think.

“I found I could put any kind of inspiration image in the algorithm and create these wonderful, unique garments,” Pritipaul said. “I can’t give the secret sauce, but we’re going to be scaling up these samples for mass production.”

Pritipaul called the entire bootcamp experience and subsequent experiments “transformative” for his career and also the whole organization.

It’s exactly what Walsh envisioned for the program. For her, A.I. is going to penetrate the entirety of their enterprise, so she wants the creators of their new tech to reflect the full company and its consumers around the world.

“I really believe A.I. can save fashion,” Walsh said. “It can optimize profitability, creativity, and sustainability all at the same time. That’s what I’m so excited about.”