At once revered and a little feared, AI is the industry’s shiny new toy and buzzword. But savvy companies agree on one thing: if AI isn’t instrumental in moving the needle, it’s not worth considering.
Such began “Betting on AI in Turbulent Times,” at Sourcing Journal’s Fall Summit featuring Kathleen Chan, founder and CEO of AI-powered “sourcing co-pilot” Calico, and James Reinhart, co-founder and CEO of secondhand marketplace ThredUp. Moderated by Meghan Hall, SJ’s business and technology reporter, the panel discussed AI use cases for both B2B and consumer-facing applications, and why it should be “viewed as an assistant, not an oracle.”
Related Stories
“When we think about AI at Calico, we’re not really looking at generative AI and making pretty pictures. We really like to think of ourselves as force multipliers,” said Chan about Calico’s suite of agentic agents that help sourcing and production managers “handle the messy middle” of manufacturing. “We’re plugged into over 150 factories in 10 different countries, and we help companies go from concept to mass production with remarkable precision, even when they’re under tremendous strain from tariffs or shifting supply chains.”
ThredUP’s AI usage has more consumer-facing applications. Earlier this year, the company used generative AI tools to rebuild its entire search functionality for shoppers. “Imagine a team of five engineers producing more in 30 days than another team did in seven years. That’s the kind of velocity we’re talking about,” said Reinhart, explaining consumers can upload their Pinterest board, Instagram handle or TikTok feed, and ThredUp tools will instantly curate a personalized shopping experience.
What’s thrilling—and frustrating—is how fast the tech is moving, and how quickly companies are testing, learning and scrapping ideas when they don’t work out.
“We’re trying to figure out is what will ultimately be the winning tools in the future, and we spend a lot of time testing stuff,” said Reinhart. “It’s a constant tension.”
For example, ThredUp’s broad customer demographic had mixed views when the company launched an AI initiative to generate images on virtual models. Newer customers appreciated the futuristic concept, but legacy users did not embrace it, said Reinhart. In the end, ThredUp shelved the idea. “We are now retraining some of the models to build something better,” he said. “That’s an example where I was very bullish on AI, and we’ve had to pull back and rethink it.”
That tug-of-war represented a larger theme—technology might dazzle, but just because you’re using AI doesn’t mean you’re using it well.
“I don’t like tech for tech’s sake,” conceded Chan. “As consumers, we’re used to seeing AI in the generative sense, but as leaders and business executives, the goal is it to see as a tool. And how do you make that impactful? Where’s the enterprise value?”
The AI revolution is also coming at a time when geopolitical supply chain disruptions and uncertain tariffs are disrupting stability, and Calico, which “doesn’t advocate” for heavy capex, has seen an uptick in users.
“Calico is so plugged into so many different factories across the world, we have that instant data set to allow folks to not only cost out what that could look like, but also very quickly move their supply lines in a matter of literally clicks,” said Chan.
AI to boost sustainability
AI also offers solutions to fashion’s worst nightmare: the glut of unsold or unworn merchandise that continues to haunt companies.
Overproduction comes from forecasting failures or execution mistakes, said Chan, adding that AI is not just about better demand predictions, but helping businesses make real-time decisions with accuracy. “I can’t control forecasting—I wish I could—but I can control with our AI, the specificity and the precision of how we execute. How do we make that process smoother? How do you bring information up front, and how do we ensure that the train keeps moving?”
Reinhart continued that thread, adding that companies need accountability. “Your incentive is always to overproduce; just like you don’t want to run out of food when you host a dinner party.” AI could change the incentives, he suggested, holding companies accountable by tracking their environmental footprint in real time.
