When Katia Walsh, SVP and chief strategy and AI officer at Levi Strauss & Co., was a young student, a teacher told her math “isn’t for girls.”
It wasn’t until years later when she was working on her Ph.D. in strategic communication at the University of Missouri-Columbia, which involved some applied math and statistics, that she discovered a love – and aptitude – for numbers.
“So many women have been locked out,” Walsh said. “I was in my twenties before I realized, wow, this is fun, I love doing it and there are so many applications.”
STEM careers are more welcoming for women and people of color, but progress requires ongoing investment.
Last year, Levi’s launched an in-house machine-learning bootcamp to help employees who don’t have a formal data science background learn about AI and develop new digital skills. More than 450 people applied for just 60 spots in the program.
Roughly two-thirds of the first class were women.
“I wouldn’t say we were surprised and we weren’t favoring women – it was a very rigorous application process,” Walsh said. “But because we made sure to encourage women and because women are equally as capable, it showed in the numbers.”
This year, around half of the class is made up of BIPOC participants.
Inclusion has been part of the culture at Levi’s since the beginning. Very fun fact: The company’s founder, Levi Strauss, who immigrated to the US from Bavaria in the mid-1800s, created a scholarship program at UC Berkeley in 1897 for aspiring designers and fashion entrepreneurs.
Twelve of the first round of scholarships were awarded to women at a time when it was not only rare for women to attend college – but more than 20 years before women won the right to vote.
Walsh spoke with AdExchanger at the Collision conference in Toronto in late June.
AdExchanger: When I mentioned to someone that I was interviewing the head of AI at Levi’s, he said to me, “Why does a jeans company have a head of AI?” So … why does a jeans company have a head of AI?
KATIA WALSH: It’s amazing how much data there is tied to everything we produce. Even a store manager is processing a ton of data, and I learned that because I spent some time working in the backroom unloading boxes, receiving orders, putting them on the shelves and actually using the replenishment app my team created to see how well it works.
I’ve worked at bigger companies from a technology perspective: Fidelity Investments, Prudential Financial, Vodafone. Fidelity is like the original fin tech. But none of those companies has the powerful global brand that Levi’s has, and I believe that brands have an incredible opportunity to make a difference in society. I can make an impact in this job.
We look at the flow of data and there are so many opportunities for automation.
Where does AI play a role at Levi’s?
Our ambition is for AI to touch everything and penetrate the entirety of the enterprise. It’s a process, though. We’ve only really been at it for the past two and a half years.
But take marketing, for example. We want to make sure our customers feel deeply connected with us, and so we have invested a lot in targeting consumers with relevant messages and making sure there is no waste in our marketing dollars.
How do you do that?
We look at three aspects: targeted email messages, the online experience across our app and the website and the buying experience.
Specifically, on the website, we personalize search so the results people see are as relevant as possible to them based on everything we know about their browsing and purchase behavior.
It’s not easy for people to search by category, especially when it comes to clothes. People may want loose fit or baggy or skinny – and these terms can be subjective. We recently started to experiment with visual search so that people can search our site by uploading pictures. It’s an experiment. We’ll decide whether we keep it based on how much consumers adopt it, but we’re always trying to provide these types of opportunities.
Levi’s is a 170-year-old brand undergoing a digital transformation. What does that entail?
People have been using the term “digital transformation” for years, but for us it’s more about how does a 170-year-old company stay relevant? We just continue to evolve and today that happens to be about digital technologies. But invention and innovation are part of our heritage. As a company, we’ve endured two world wars and multiple pandemics, including the Spanish flu.
Why is diversity such an important part of this transformation?
It’s about relevance. A brand should reflect its consumer base and its employee base. The world is diverse and so we need to be diverse in order to stay relevant. It’s a business imperative that also applies to technology and particularly artificial intelligence.
There is an issue with bias in algorithms that becomes pronounced when only limited data is available or if it’s based on a certain agenda. We want to make sure we don’t perpetuate bias, of course, but we also want to minimize it. I don’t think it’s ever possible to fully eliminate bias – the world around us is inherently subject to bias – but it’s imperative that we do our best as both human beings and professionals.
Can you share a few practical examples of how Levi’s is applying AI to the business today?
Everything in life is a network. There are physical networks that connect us through our phones, but also human ones. When I first started at Levi’s I remember thinking, “I wonder if our products are a network, too?”
We did analysis starting in China – we’ve since applied this all over the world with other products – to understand how our products relate to each other and which are complementary so we can make recommendations that go beyond personalization. When people buy certain products, for example, it boosts the purchase of others, like a black T-shirt that leads people to buy the same one in white and other colors. We also learned that some products cannibalize each other.
We’ve taken what we’ve learned and used it on our site, in our app and in our stores. None of this is theoretical.
This interview has been edited and condensed.
