It turns out the revolution in artificial intelligence requires a lot of good old-fashioned thought by living, breathing humans.
The hype around the lightning-fast technology and its real world potential is consuming vast quantities of chief executive officers’ time, brain power — and millions or billions of dollars of corporate resources, stoking worries of an economic bubble that could deflate quickly.
It’s impossible to tease out exactly how much fashion is spending on AI, but the tech giants on the frontier are pumping endless funds into the revolution. The leading players are expected to spend $5.3 trillion on AI and data centers by 2030, according to Goldman Sachs.
The rest of the corporate universe is still trying to figure out how and when to really jump in.
A WWD study of the annual reports of 17 U.S. retail and fashion companies that project capital expenditures found “information technology” to be a common thread. And while it appears retailers are still prioritizing their stores, the bread and butter of their businesses, technology is often listed as the next category down on the spending list. Omnichannel and supply chain spending also showed up repeatedly.
The biggest retail spender by far is Walmart Inc., which plans to devote a blockbuster $25 billion to $27 billion to capex this year after spending $25.6 billion in 2025. But the rest of the industry is hitting the accelerator hard. Total expenditures by the other 16 retailers in the study are seen jumping 25 percent to $13.9 billion, at the midpoint of 2026 estimates.
Dick’s Sporting Goods Inc., which plans to put $1.5 billion, or an industry-leading 8.5 percent of its revenues, toward capex this year, describes its spending as, “ongoing investments in our supply chain and technology…investments in technology to enhance our store fulfillment, in-store pickup and other foundational capabilities to drive efficiencies.”
Fashion and retail companies are known not to be high-tech leaders, but fast followers — or at least eventual followers — when it comes to the latest tech.
But, in this, the industry is not alone. The corporate world — and society — are still taking the measure of AI and trying to figure out how best to put it to work.
As companies look to integrate AI into their systems, they’re largely using a “broad, but shallow” approach, said Rebecca Homkes, a consultant and high-growth strategy specialist.
“The majority of folks in the majority of teams are still using AI for individual productivity,” Homkes said. “Organization-wide or even cross-team gains that can be measured are incredibly limited.”
So an AI leap forward in, say, sampling, might help the design team churn out twice as many new looks, but if the process to choose styles doesn’t keep up, opportunities are lost. Or styles might get to the stores faster, but marketing might struggle to engage the right customers.
“Within the pieces, tasks are more productive, but we’ve actually just introduced all of these new bottlenecks in the organizations,” Homkes said. “There’s still so many protocols in place for human oversight of AI, which is for a good reason because AI can still make mistakes. But until we step back and redesign work for a world of AI, gains are going to be incredibly limited.”
To really redesign work, it’s going to take a pretty big step back.
“You cannot make linear plans against exponential change,” she said. “We’re very used to, from the digital transformation world, here’s my three-month milestone-based way of how I’m going to change. By the time you can be halfway through that workflow, the AI tools have changed in a way that this no longer makes sense.
“It’s not just that technology’s moving very quickly, it’s that how we govern and run change management inside organizations has not adapted at the pace of this technology,” she said.
Homkes recommends companies switch to “parallel pathing” where key elements of a change are worked on at the same time, rather than sequentially.
“There’s a governance plank, a capability plank, a tech platform plank, and a linkage to outcomes plank,” she said. “They all need to be moving at the same time in four pathways, which is a little bit uncomfortable because I want all the governance in place before I deploy, but that isn’t really practical because the governance needs to keep adapting as well.”
Newer companies have the ability to change quickly, while established names have an existing customer base and other strengths to play off of.
But Homkes said the race to really use AI was not about new versus old.
“It’s speed of adaptability that really matters,” she said. “Agility is based on methodology. I keep banging my head on this. Agility is based on methodology. The broader discussion in the industry is about size of a company. ‘Oh, I need to eliminate all of my managers because with AI I don’t need them and I want to be smaller and more agile.’ The agility you have as an organization is not directly correlated to number of employees. Agility is based on methodology.”
This thread — that the switch to AI involves not just a lot of fancy new technology, but new ways to approaching the human side of work — is one that comes up often when speaking to experts in the area.
Greg Portell, lead partner in global markets at Kearney, underscored the importance of not just plugging AI into a process, but of figuring out where and how people need to be logged in.
“Companies have known and had really good decision intelligence for a while now,” Portell said. “The number of companies I’ve worked with that have excellent planning software and then, when the moment comes to make the decision, they opt for something different because a planner knows better. There are a lot of companies in that boat.
“If I’m an apparel company and I know that avocado green is going to be the color of the spring, does that help me order more product? Does it help me order better sizing? You have to figure out AI, but you have to pick the right business case.”
Portell suggests that companies, faced with a torrent of change, work not just on new technologies, but on their own ability to funnel that change.
“After AI, we’re going to do quantum [computing],” he said. “And after quantum, we’re going to do something else. The change is coming cataclysmically faster than companies can adapt. And so that’s why the emphasis — as we skew it on the technology — is probably a little bit misplaced versus the management discipline to act on that technology.”
But one way or another, almost everybody seems to agree that this is a wave of change that can’t be avoided.
“AI has to be integrated into the business,” Portell said. “It’d be like a business saying we’re not going to do email or we’re not going to use spreadsheets. It is such a core part of the way companies work and the way society works going forward that you have to be involved.”
That paints a picture of a business landscape that’s not only evolving at light speed, but one where almost every function within a company could be radically changed — and then need to be changed again.
“Most CEOs still scratch their heads and wonder how much money they should devote to it,” said Achim Berg, managing director of FashionSights. “They also realize that AI is not free anymore, but that it really comes with a cost because if everybody’s now building agents, you have to manage that and you have to provide budgets for that.”
There’s been a lot of talk about brands using AI to work on the consumer-facing front end of the business through curation and marketing, connecting consumers with the right look at the right time, nanosecond by nanosecond.
“But I think the larger part of the opportunities are at the back end, which is actually the part that was neglected for a very, very long period because everybody was e-commerce, social media — everybody was so focused at the front end,” Berg said. “This industry is ridiculously old-school. When you think about nine- to 12-month lead times, product cycles, buying and sourcing processes, most companies haven’t really changed those in 15 or 20 years.”
CEOs have only so long to scratch their heads and figure out where to really push.
“There’s no way to meet your board and not to talk about AI,” Berg said. “The big challenge is, ‘How do you develop a plan that also makes the return on investment tangible?’ That is the piece that is incredibly difficult and pretty vague.”
Levi Strauss & Co., one of American fashion’s oldest companies, has been among the most ready to take on the technology.
Michelle Gass, president and CEO, has been stressing the importance of incorporating AI into the company as she looks to complete the firm’s years-long transformation from a men’s wholesale bottoms business to a head-to-toe denim retailer.
“We’re a growth company,” Gass told WWD. “We have to support our growth and we want to drive even more profitable growth. And so how do we leverage [selling, general and administrative expenses], keep our costs under control? We see AI as a big enabler to that, really helping with the capability and capacity of our teams. We’re seeing adoption across the board to automate legacy processes.”
That translates into a lot more players in the game.
“We’re up to around 1,000 agents that are being deployed to do automation,” Gass said. “We have that happening in supply chain, in planning, in the orders that we get from our wholesale customers. We have many, many examples now.
“We’ve been in the midst, we’re almost done, of this big [enterprise resource planning] transformation that has solidified our foundation around data,” she said. “That’s been crucial to now layer on all the AI capabilities. So I’m feeling quite good.”
How many agents should be flitting around a fashion company is an open question. A 1,000? 10,000? A million? Does every customer get their own dedicated brand agent?
But as the artificial IQ expands, again, it seems the grey matter matters just as much.
“The CEOs are pushing, use AI, use AI, use AI,” said Alex Yaseen, founder and CEO of Parabola, which builds AI agents for operations and finance teams and works with Skims, Faherty and others. “But a lot of these finance ops teams are finding that they haven’t had success. So they know they need to and they’ve been struggling.”
But that, like everything else in the world of AI, is changing fast.
“It’s really just been much more recently where you can actually see some teams who truly buckle down, focus on here’s the way-in-the-weeds details of exactly how this process works and ironing out all those things,” Yaseen said. “It’s the people issues. They’re human change management issues.”
When doing things is easier, the focus shifts to deciding what to do, leaving people not just in the loop, but deciding what the loop is — at least for now.
“Even pre-AI writing code was easy,” Yaseen said, pointing to an area where AI has brought big changes. “There’s a learning curve to it or whatever, but knowing the details of the problem is the complicated thing. And now with AI, that’s extra true.”
Parabola, which was founded in 2015, helps companies not just use AI to make decisions, but provides a kind of dashboard to track and understand how those decisions are made and ensure the inputs are accurate.
To custom fit that for each company, Parabola has what Yaseen described as a team of “automation engineers” who spend time with clients to understand what they need.
There’s a lot of angst about how AI will ultimately impact the labor market, but much of it still exists in the fuzzy future.
But today, this is the kind of job AI is creating.
So who is Yaseen looking to hire as an automation engineer?
“There’s some intellectual horsepower, intellectual curiosity,” he said. “They’re able to engage really deeply and learn things quickly, but almost more importantly, almost like an instinct or a compulsion to go deep on things where you have that intellectual curiosity and you can’t not ask that next question about how does that work and why?”
That sounds a lot like the mental gymnastics CEOs everywhere are going through now.