Yahoo Finance

“That comes down to predicting our customers’ behavior, not just demand, but behavior, and the tool helps us with that as well,” he said.

2) Optimizing inventory planning will bridge gaps in the supply chain

Having solid systems in place to forecast demand can make tracking inventory needs, and subsequently planning based on those needs, easier for retailers of all sizes.

That benefit can come from making better bets on orders that will ship six to nine months after they’re placed, or from more accurate short-term data needed to make decisions about replenishment of specific products.

Tailored Brands has already begun to see the benefits of data and AI on short-term inventory decisioning. In its rental business, the company previously saw a fair amount of drift—that is to say, items rented at one store ending up at a distribution store across the country after being returned by the consumer. That caused a logistics headache for the company as it fought to meet the expectations of the next consumer in line to rent the same item.

AI has helped the company sense and respond to such shifts and prepare inventory as needed, particularly in key markets. In turn, it has seen fewer dollars spent on emergency shipping and fewer unhappy customers forced to make a substitution for an out-of-stock item in the lead-up period to their rental.

“If you get that [balance] right, you spend a lot less money shipping out of market or doing transfer shipments,” Vifquain said on stage.

3) Localizing relevant inventory can be a form of personalization

As demand forecasting capabilities proliferate, inventory planners stand to benefit—both from the historical data aggregated by AI and external data, like how size curves have shifted—to create stronger assortments and allocate more accurately.

Tailored Brands has already started to see the results of that type of technology in action; for its rental businesses, it has been able to better understand the average shifts to size curves in each of the U.S. states—and how they could be affecting men’s formalwear rentals.

Vifquain said, prior to implementing AI systems like these, Tailored Brands relied heavily on anecdotes from in-store associates, who may have asked for more of a specific size without concrete data. Now, AI systems like Invent’s can show which sizes are being rented most, and where the local or state size curve could be headed, at the click of a button.

“The beauty of this [type of AI system] is it’s actually adjusting dynamically this size profile based on demand, and saying, let’s start to lean more towards small or large for this store going forward, because that’s what the demand is asking for. So it’s distorting the assortment and reducing inventory. That’s a pretty powerful combination,” Vifquain said.

That kind of precise, localized understanding of consumers’ needs, patterns and desires can also allow brands and retailers to stock a smaller quantity of inventory while still turning higher profit margins; without excess inventory clogging systems and space up unnecessarily, consumers can cut through the noise and receive unfettered access to what they truly seek, the presenters noted. That, in turn, can make the customer experience feel more personalized, even though the inventory allocations have been decided upon based on data around a geographic segment or trend.

4) Agentic AI could be on the rise

Throughout NRF 2025, technology providers hyped the incoming power of agentic AI, which is a type of AI that effectively enables models to complete tasks or make independent decisions based on a slew of data and inputs.

For instance, a planner might direct an agent to determine how many units of a SKU should be allocated to two of its stores—one in Los Angeles and one in New York City. From there, the agent can pull data on the historical performance of each store, analyze weather patterns that could influence purchasing decisions and use social media data to determine what’s trending on the West Coast versus the East Coast.

Kok noted that there are two main types of agents today: those that run nearly entirely in the backend, streamlining systems and monitoring decisions that have already been made, and those that interact more fluidly with human users to advise on issues like planning. The latter type of agent helps aggregate and pull insights from complex, murky data sets—sometimes with millions of data points—to glean insights into how the business could add value.

Agentic AI, Kok noted, is likely to help eliminate some of the most mundane tasks those dealing with inventory, merchandising and pricing deal with today.

“People want to make better decisions, using better tools, and they will have a better impact on the business with a lighter and more fun workflow,” Kok told Sourcing Journal.

5) Data democratization will be key to powering the systems that make the magic happen

Brands and retailers hoping to gain the maximum benefit from the type of AI systems Kok and Vifquain described throughout the session will only be able to do so if they break down internal data silos, Kok said.

Much of the talk about training AI systems centers around having clean and complete data. But Kok said, if data is relevant and mostly accurate, every little bit helps companies that are working to improve their supply chain processes with technology.

Moreover, AI can help detect where inventory data is partly incorrect or incomplete. Kok said that the important part of getting started with an impactful AI journey is just that—to get started, sooner than later.

“Perfect is the enemy of better,” he said.

In some cases, important data could be residing within another function of the company, preventing it from being used to train or query AI-based systems. Unless that data becomes more widely available internally, it’s not doing the enterprise any good, Kok said.

“You can have a gold mine, [but if] it’s 10,000 kilometers under a mountain, it’s not valuable,” he told Sourcing Journal.