The saying goes “Too much, too little, just right” and Goldilocks and her diridge. Balance is also important for inventory management. Part of the supply chain is to analyze what consumers buy and make sure their products are in stock at the right place and at the right time.
Excessive inventory can lead to markdowns and expired products, but too few products can lead to shortages that affect the retailer's brand image, customer satisfaction and revenue.
To prevent stock from running out, big box retailers like Target, Home Depot and Walmart use AI to predict when product volumes will decrease. As a result, Target inventory availability has improved every year for the past four years, Target's executive vice president and chief information and product officer, Prat Vemana, told Business Insider.
Rather than responding to changes in conditions, AI helps retailers actively adjust before a suspension strike, according to Ajoy Krishnamoorthy, CEO of inventory management platform CIN7. This is particularly important today, Krishnamoorthy added that factors such as consumer behavior, inflation and trade policies have constant influence on the supply chain.
“AI thrives in this environment,” Krishnamoorthy said.
From old methods to AI
Traditionally, companies source inventory, manage logistics, analyze consumer behavior and analyze consumer behavior in silos, according to Vidya Mani, an associate professor of business management focused on technology and operations management at the University of Virginia. The team does individual research, bring it up and develop a strategy and implement it.
“We don't have that time anymore,” Mani said. “By the time you finished it, the world would have changed to you.”
According to Vemana, retailers relied on software-based applications before targets could use AI to predict stockouts. In fact, Target said in a blog post that he couldn't catch half of his previously out-of-stock products.
Target changed the way it manages inventory in 2023. Inventory Ledger implementation uses AI to track inventory changes across stores and to use AI to predict that products may run out of stock “even before they become apparent to team members or systems.”
Today, Target uses AI-driven inventory management for over 40% of its product assortment. That's more than doubled when it first began two years ago, Vemana says.
With inventory ledgers, the algorithm draws in data such as supply lead times, transportation costs, current inventory, and consumer demand. Targets use both types, as some models are more accurate in the categories you buy frequently, while others are more suitable for discretionary purchases and clearance items. There are also models that detect items in the store but are in the wrong aisles and shelves, Vemana said.
Target has a demand forecasting tool that says, “make billions of forecasts each week about the number of items needed in stores and online.” Together, all of these technologies guide employee decisions on when, where and how to sort products to restock.
“Combining traditional software and AI allows us to make smarter, faster decisions about inventory management and make our stores more consistently stocked,” Vemana said.
ABC School Supplies, one of CIN7's customers, also uses AI to access real-time sales data, potential stock shortages and supplier lead times, so “we can “actively sort and avoid costly gaps in supply,” says Krishnamoorthy.
AI-driven inventory management shows a major change from what ABC academic products have done in the past. Krishnamoorthy said he was copying and pasting website orders into the system, creating pick-and-pack lists on paper, walking the physical list to the warehouse, and manually updating inventory.
Home Depot also takes an approach to AI-based inventory. In 2023, retailers deployed a machine learning-based app called Sidekick. This leads the clerk to restock the shelves, restock the shelves overhead and find products, among other features.
“It helps to ensure that products are on shelves for our customers, and manages the accuracy of the hands we supply to our refill and sales platform,” a spokesman for Home Depot said.
AI's predictive power over inventory planning
Krishnamoorthy said in retail, “AI is exploding” as businesses move from static plans to forecast demand and dynamic forecasts that prevent stockouts.
AI allows businesses to avoid stocking out at specific store locations and peak times and get more detailed inventory data. Mani showed examples of cosmetics and how different stores can change product needs based on consumer demographics.
“AI can grasp the patterns of baskets that are frequently purchased in these different clusters,” Mani said. “There's no need to give that contextual information.”
Target is also working on technology to predict what seasonal items colours and sizes will be sold, so they can stock these items to “meet local demands.”
At Walmart, AI-based inventory management systems use algorithms to ensure that warm state stores stock adequate amounts of pool toys and cold states. If a particular item is not sold on the East Coast but is popping out of the Midwest shelves, the algorithm flag will in that pattern allow Walmart to relocate the inventory.
As retailers continue to develop and deploy AI tools, Mani predicts technology use cases will advance over the next decade.
Mani said in two or three years, AI is likely to flag out inventory without anyone who doesn't have to step into the store to check. Five years later, AI could automatically sort the inventory if the algorithm detects low inventory. Also, in 10 years, AI will understand how macroeconomic events (such as inflation) will change future patterns of consumer buying behavior and adjust inventory plans accordingly.
“It doesn't respond to the situation, it gives it to your rulemaking,” Mani said. “You live in a different world.”
