If there was one clear message at NRF 2025, it was that AI is over the peak of the hype cycle. Multiple sessions featured retailers and software vendors presenting practical use cases where generative and predictive AI tools are delivering tangible results. In an earlier Insights from NRF 2025 blog, we provided examples to show how retailers are tapping into AI to support frontend operations. Here, we focus on backend operations, exploring applications that are helping retailers improve inventory management, pricing strategies, store planning and supply chain operations.
AI for inventory management
In the session “AI simplified: A modern approach to profit-optimized retail planning,” Brian Kilcourse, Managing Partner at RSR Research, shared statistics from an upcoming report confirming that 45% of retailers say their top inventory challenge is not being able to accurately forecast demand. As a result, retailers run out of popular products and have surpluses of unpopular products. They identified inventory allocation as the second biggest challenge — figuring out how to avoid stores ending up with too many of a product while other locations end up out of stock of the same product.
By analyzing internal data and external data, AI tools can quickly see patterns and make recommendations that help retail planners make better decisions. Of course, because machine learning is integrated into the tools, feeding them new data improves their accuracy over time.
Jennifer Chaffee, Senior Manager of Merchandising IT at Mattress Firm, who also spoke at the session, said that her company teamed up with invent.ai specifically to solve its inventory issues. She explained that with 2,300 stores and 63 distribution centers, the forecasting and store planning process is difficult and time-consuming. “Beyond forecast accuracy, it’s about getting the product in the right location where it has the chance to be most profitable.”
It comes down to financial ROI. Improving forecast accuracy will improve our availability of products, so customers can get their first product choice. This will reduce our outlet inventory because more customers will be happy. We’ll be able to optimize discount depth and have better margins. All of this was modeled in advance along with the cost savings in our supply chain.
Senior Manager of Merchandising IT, Mattress Firm, NRF 2025
invent.ai spent 10 years refining its AI solution which incorporates predictive, generative and agentic AI actions. Gurhan Kok, founder and CEO, who shared the stage with Brian and Jennifer, said it connects internal data sources like sales and inventory with external sources such as weather, local events and social media factors to help retailers refine assortments, optimize inventory and streamline pricing. It can even work with incomplete data, refining its analysis as it gains more and better data. “Our goal is to help retailers reduce excess inventory while ensuring availability so retailers can sell more products in the right places.”
He pointed to US retailer Academy Sports as an early success story. With 298 stores and over one million SKUs, inventory management was challenging. In 2018, the company launched a pilot program for inventory optimization and replenishment at the store level. “In the A/B test, the stores that used AI improved inventory turns by 8.6% and recognized 420bps — that’s a nice chunk of money to find under your pillow within months. “
Gurhan advised retailers in the audience to take a “data to dollars” approach in entering the AI game, choosing a use case based on how fast it can turn their data into dollars.
Some use cases can take a long time to deliver results. Look for something that's a fast data-to-dollar exercise. You want to unlock value and ROI that can fund a future project. AI tools can make an impact on things like inventory, promotions and dynamic pricing within months so that’s the best way to go.
Founder and CEO, invent.ai
AI for pricing strategies
While the vendor booths in the NRF Expo offering AI-driven electronic shelving technology that enables retailers to deploy dynamic pricing in real-time were consistently busy, none of the sessions featured this use case. However, in many of the AI-focused sessions, retailers touched on how they see the technology supporting pricing in the future. For example, in the session, “10 AI trends shaping 2025 and beyond,” Scott Vifquain, Chief Technology Officer at Tailored Brands, which owns Jos. A Banks, Men's Wearhouse and K&G Fashion, shared that the company is exploring the idea of using AI to support the hyper-localization of pricing. He cautioned that they’ve learned this isn’t as easy as it sounds as the prices the customer sees online need to match those they find in store.
In the session, “Applied AI: Where demand planning technology and markdown optimization meet,” Juan Hernandez, Director of Omni Channel Pricing at Dick’s Sporting Goods, presented a tangible pricing use case that is working for the brand, markdown management. The 850-store brand is leveraging an AI-powered tool developed by Oracle Retail Labs to support its markdown strategy.
Juan shared that the company constantly adjusts prices, creating around 11 million SKU combinations every week. He said the new tools are eliminating a lot of the complexity, making it easier and faster for his team to plan and manage weekly discounts and sales.
AI for store planning and supply chain management
For the NRF 2025 opening day keynote, Azita Martin, VP of AI and GM for Retail, CPG and QSR at Nvidia sat down with NRF Chairman and Walmart US President & CEO, John Furner to explore, “AI in the real world.” After briefly reviewing how Nvidia’s technology is helping the world’s largest retailer improve inventory management, the discussion turned to how Nvidia sees AI disrupting supply chain management in retail.
Azita introduced a concept pioneered by Nvidia called physics AI, which is AI that understands the laws of physics, i.e. the weight and volume of things, which allows retailers to create digital twins of anything — physically accurate representations of anything from individual products to stores and distribution centers. The company worked with Lowes to create twins for all 1,700 of its stores, enabling the home improvements retailer to analyze how people and objects behave within the spaces and explore potential improvements without disrupting operations.”They update several times a day with operational and inventory data, and have been able to optimize how customers are shopping their stores and ultimately improve their sales and revenue.
With 80% of revenue still occurring in physical stores, being able to simulate different layouts to optimize operations before you make a capital investment is exciting and extremely important.
VP of AI and GM for Retail, CPG and QSR, Nvidia
To give an example of physics AI in action, Azita showed a video featuring a digital twin of a distribution center updating in real-time. When an incident occurs, in this case, a pallet of boxes falling off a shelf, an AI-enabled camera sensor catches it and sends a message to another AI model to reroute an oncoming forklift. It also sends an alert to warehouse management and provides an analysis of how the incident occurred and how it can be avoided in the future.
Conclusion
Retailers deal with multiple uncertainties on a daily, monthly and yearly basis, along with the pressures of generating higher profits and lower costs. By processing and analyzing information quickly and accurately, AI tools make it easier for retailers to make the best decisions for their business. As Gurhan explained, “What it comes down to is that, as humans, we're just not able to take maximum advantage of all the data available to us.”
However, he doesn’t see a future where AI replaces humans. It’s simply an enabler. “I don't think anybody gets up in the morning and says, ‘I have to get some more AI into my company.’ The objective is: How can I make my company more profitable and leaner? What tools will help us make better decisions and make things easier and happier for our employees?”
To explore the biggest trends, technologies and opportunities shaping the future, download our white paper, Reimagining Retail in 2025: How retailers are adapting, evolving and thriving in a changing world.
Anita J. Temple is the Corporate Journalist at commercetools. She was a fashion editor at Women’s Wear Daily (WWD) and W Magazine before launching a career as a freelance writer and creative producer. She has written content and worked on a wide range of marketing projects for companies including Dreamworks, Walmart, Coca-Cola, Verizon, and Adidas.