(©KarolaG - canva.com)

In a keynote session kicking off the National Retail Federation’s annual ‘Big Show’, Fanatics founder Michael Rubin told the story of how the relentless pursuit of innovation turned a small e-commerce platform into a multi-billion-dollar enterprise. But if delegates to NRF 2026 started the show with a clear picture of how disruptors generate success in retail, many will have left the three-day event with their heads spinning. With hundreds of solutions and technologies on display, they could be forgiven for wondering how to wade through the hype to harness the biggest disruptor of them all - artificial intelligence.

The temptation is to rush in and get ahead of the game. This should be avoided. Solutions may promise to solve every forecasting, planning, and logistical problem in the retail supply chain. But AI-driven transformation starts long before buying the software.

Of course, there’s nothing new about the use of AI in retail. For more than two decades, machine learning capabilities have been powering forecasting capabilities (since 2003 in the case of our company), helping retailers get the right inventory to the right place, in the right amount, at the right time.

But while these technologies once worked behind the scenes, generative AI changed all that, placing them firmly in the hands of users. By interacting with AI agents, retailers can now determine how many blueberries to order to prevent stockouts or which of their furniture outlets should carry chairs and tables but not beds or ottomans. With intelligent agents making business decisions on behalf of users, humans and machines now work together directly. This is a new world.

Yet despite high rates of generative AI piloting across retail value chains, a 2024 McKinsey report found few companies were yet realizing the technology’s full potential at scale. Barriers range from lack of technical capabilities to implementation expenses. But part of the problem may be that retailers are not asking themselves the right questions at the outset.

Three of the most important are - what can AI agents do—and what can’t they do? How can humans be equipped to work effectively with machines? And most importantly, which AI solutions can be tied to specific business outcomes? For companies without answers, AI will be a box of expensive tools that ends up sitting on the shelf – or worse hurting the business more than helping. So, what should retailers be thinking about?

A task for every agent

In retail, it’s hard not to find a use case for AI. Here’s one example. In footwear, you can tell your agent to update the forecast for next month’s tennis shoe sales based on a potential 2% lift because of a promotion. The agent reviews the forecast, presents the results, and, from order entry to transportation, makes the necessary adjustments. In fact, at every step in the supply chain, AI technologies have astonishing power to increase efficiency and inform decisions by making sense of vast amounts of data at speeds humans would find impossible.

Take merchandising. For retailers with thousands of outlets, some stores might be just a few miles apart but need very different product assortments. AI can recommend what to carry and where at a granular level. Once inventory decisions are made, AI supports warehouse placement, slotting inventory to optimize collection, storage, and replenishment. And when it comes to transportation, AI will optimize truck and ship movements for everything from cost to efficiency. At the store, AI can design work schedules based on the number of employees available, seasonal sales forecasts, and customer traffic patterns.

Of course, deploying AI solutions across every part of the supply chain may not always be affordable or feasible. This makes it essential to prioritize. And the first step is to pause to review the options and assess how and where AI tools can benefit the business, what people skills will be needed, and what underlying data must be in place.

The path to business alignment

Many of the products on offer today are a solution looking for a problem. But there’s no point, for example, in prioritizing buying an AI solution to help build out a product line if your design team is really solid – that money may be better spent elsewhere. So before making any AI investment, a key question is this: what problem are you trying to solve?

Numbers can help. For every hole on a shelf, you can calculate the revenue lost when customers enter the store and find the products aren’t there to buy. So if your stock replenishment is weak, that could be an area of focus. Whether it’s in-stock levels or OTIF (on-time, in-full) delivery rates, industry standards can inform decisions on which AI solutions to deploy and where to deploy them to maximize return on investment.

But if AI technologies are to transform retail, humans will power that transformation. This makes investments in people as critical as those in technology. First, existing teams should be engaged early in purchasing and implementation decisions. For new hires, onboarding needs to be redesigned. While entry-level associates once learned from the experts in their function, training must now focus on ensuring they fully understand the capabilities of the AI tools they’ll use.

When it comes to new hires, this is not a numbers game. It takes a relatively small team to scale up AI adoption by designing the strategy and, over time, infusing AI capabilities throughout the enterprise. Meanwhile, as retail supply chains evolve from linear functions into digitally connected ecosystems, a senior leader should ensure that cross-functional AI strategies align. If your warehouse agent is optimizing for labor costs but the inventory agent is optimizing in-stock, all you will have done is automate siloes. Someone needs to make sure the intelligent agents in each function are working together—orchestrating for optimal outcomes—not fighting each other.

Meanwhile, any AI tool is only as good as the data available to it – agents will struggle to operate across data silos. This requires bringing together all data across the supply chain. Whether it’s accurate cross-functional metrics, information on end-to-end workflows, or global real-time visibility into disruptions, AI planning and forecasting requires the right data and right numbers to crunch.

Where’s the business case?

These days, AI dominates industry conversations—it was certainly prominent onstage and in the networking sessions at NRF 26. And on everyone’s lips is the question of whether or when the AI bubble will burst. Certainly, if retailers’ AI investments fail to gain user buy-in, are based on incomplete, inaccessible or unreliable data sets, or do not align with a business case, they could get into trouble.

However, when used appropriately, AI will fundamentally change retail. And the leaders won’t be those spending the most money. They will be those planning carefully for AI adoption, engaging users early, rethinking how they build human capabilities, putting the right data in place, and ensuring end-to-end visibility through single data models on networked platforms. But above all, leaders in the age of AI will be those with a clear picture of problems they want to solve and which tools they can use to solve them. This will be the blueprint for enduring success.