Aaron Ricadela | Content Strategist | Sep 5, 2023
Recent changes at grocery stores and supermarkets are among the most profound in the retail sector. Since the pandemic started, workers have fled to other sectors, shoppers have completed more of their grocery buying online, and grocers have had to rethink their merchandise assortments and how long shoppers want to stay in their stores. Even as grocery retailers have shifted part of their business to their own websites and apps, they’ve been challenged by volume discounters, big-box stores, and specialized online delivery services.
Artificial intelligence software can help grocery retailers address those challenges as well as optimize prices, inventory levels, and the shelf placement of goods. To be sure, the extent to which the industry uses AI has been overblown at times, as have the technology’s capabilities, but it’s starting to be rolled out for select applications.
Supermarket chains and regional grocers are adding automated shelf scanning, smart shopping carts, automated payment systems, and other AI-based technologies. Grocers are testing mobile apps that can personalize shopping lists based on dietary preferences or what a shopper wants to cook that week, as well as software that lets store managers see how products sell based on their aisle placement. Computer vision systems keep track of inventory on shelves, and other AI systems let consumers pay automatically, without going to a checkout line, by tracking what’s in their carts. AI software also helps grocers experiment with price changes to maximize profits and see the effects of price shifts on products frequently bought together—such as whether lower prices on corn chips would boost salsa sales.
To fuel AI-based decisions, grocery retailers collect huge amounts of data, mainly from their point-of-sale systems, ecommerce sites, loyalty programs, and in-store cameras, as well as information on the weather, nutrition, and demographics from external sources. They store it in data warehouse software for statistical analysis.
High transaction volumes involving a large number and variety of products mean grocery retailers generally have more data available for AI analysis than most other retailers, letting savvy store chains make decisions with statistical confidence.
AI software helps food retailers draw insights from huge amounts of data and show tens of thousands of metrics and key performance indicators tailored for their industry. The applications can help grocers experiment with prices before a broader rollout, provide real-time reports on shelf stock, robotically pick goods for ecommerce orders at “micro-fulfillment centers” at the back of stores, and perform many other tasks. The software also helps supermarkets see signals in demand and allocate inventory across their stores to where it’s needed most—depending on demographics, the weather, nutritional trends, and other factors—helping them avoid both stockouts and excess inventory.
Unlike products in other retail categories, many grocery goods are perishable, putting a priority on just-in-time stocking decisions that AI algorithms can help inform. AI software also can help grocers make pricing and shelf placement decisions about individual products and where to place goods, often bought together or substituted for each other. Grocery retailers also use AI to create “heat maps” of spots in their stores where customers tend to walk and dwell most often, informing their planograms on how they lay out shelves, endcaps, and other displays to make it easier for customers to find products (and help the stores maximize revenue). Another AI application helps retailers see customers’ “decision trees”—graphs of how customers make product choices, and the order and importance of various product attributes. An algorithm weighs how shoppers evaluate assortments and predicts what they might buy if their desired product isn’t available. That analysis helps supermarkets optimize planning and ordering—so customers don’t leave a store empty-handed.
Supermarket chains collect data for AI analysis at the store level, but the data processing happens in their data centers or those of cloud service providers. The user of the AI-based software is often an inventory analyst who examines business data across the store chain. AI most often pulls the data from a centralized merchandizing system.
Grocery stores are putting sales history, pricing, and customer attribute data to work in new ways via a variety of AI-based software and systems. Here are six of the top applications for that technology.
Cashierless checkout systems like those at Amazon Go stores or those supplied by Grabango and other vendors help shorten store visits, improving the customer grocery experience. Israeli startup Trigo has rolled out a cashierless supermarket in Munich, Germany, for the Rewe chain and Aldi Nord. Meanwhile, chatbots help automate online ordering. For example, Walmart’s technology incubation unit has developed mobile phone software that lets consumers at home speak or text items they want placed in their virtual shopping carts and schedule deliveries. Instacart, which lets consumers shop across a variety of grocery store brands, is adding AI-based chatbot technology to let shoppers ask for recipe ideas or about ingredients.
Grocery retailers use AI software for food quality inspection. For instance, aisle-roaming robotic scanners see how quickly meat or vegetables sell to gauge their freshness. AI algorithms factor in the weather, local events, and other external data points to guide grocer purchasing. AI systems can also inspect incoming pallets of goods for damage.
AI systems can help retailers look across their stores’ product assortments, analyze historical sales, and find signals to prevent stockouts. For example, Simbe’s Tally 3.0 robot roams aisles to look for out-of-stock or out-of-place products or assortments that don’t adhere to stores’ plans. Simbe and other systems can also generate heat maps of spots on shelves generating the briskest sales. Walmart is outfitting its stores with cameras, sensors, interactive displays, and servers to identify when perishable goods and other items run low on shelves, triggering notifications in internal applications that alert associates to restock. Grocery retailers can also look at factors such as weather or eating trends affecting demand at stores spread over a wide area.
Food retailers use AI software to personalize promotions across various ways customers communicate. For example, knowing that a customer frequents a certain store each week, likely has a new baby based on purchases, and responds well to certain coupons via text message can help retailers create effective electronic coupons. Personalized promotions are considered one of the most promising grocery s use cases for advanced data analysis techniques, which yield fresh views on the effect of promotions and the sales effect that certain promotions have on other products. According to consultancy McKinsey, when fielded with precision, promotions can lift sales by 4% to 8% and boost operating profit by 2% to 3%.
AI algorithms can detect changes in shopper buying patterns and suggest prices for goods that maximize sales. Such analyses factor in price history, inventory levels, competitor prices, supplier costs, and other data points. Retailers can also use AI to perform an “affinity analysis” that suggests the best prices for complementary items, such as coffee and creamer or chips and salsa. In addition, AI analyses can reveal the “demand transference” of substitute goods.
AI-based computer vision applications can reduce theft by identifying shoppers who pocket items from shelves and cashiers who don’t scan each item at checkout. Stores can position cameras over self-checkout lanes to determine whether customers intentionally pass products over a scanner bed without registering their barcodes.
An estimated 30% or more of the world’s food supply goes to waste at both the retail level and by consumers. AI software can improve forecasting so stores don’t overorder perishable foods that may later go to waste. Instead of reordering foodstuffs based on intuition or historical parameters, AI software lets grocers analyze data on weather forecasts, local events, nutritional trends, and from other sources to predict demand and supply their stores more accurately.
Israeli startup Wasteless’s machine learning software lets supermarkets implement dynamic pricing so perishable items get cheaper closer to expiration dates. A government-funded project in Germany uses AI algorithms to help food producers generate products such as ground meat with longer expiration dates, minimizing the energy used in the mixing process.
Shoppers roaming the aisles of a modern grocery store are apt to encounter more technology changes than ever since the spread of barcode scanners 40 years ago. Robots might patrol floors, checking shelf stock and cuing workers to replenish waning supplies. Electronic price tags change throughout the week so software can analyze how variations throughout a chain perform.
Some supermarkets have also built back-room automated warehouses that use computer vision and robotics to prepare goods for ecommerce orders. Self-checkout is moving past handheld scanners to systems that detect the items a customer has placed in their cart, then charge their account when they leave the store.
AI software can analyze millions of price variations across stores depending on demand, season, day of the week, promotions, economic factors, and other changes and then recommend prices for items to maximize profit. Supermarkets, in turn, can replicate prices that work across cities or regions.
Supermarkets use AI to look for patterns, such as changes in shopper habits that indicate shifting lifestyles and how that person responds to offers. Taken together, stores offer tailored promotions.
So-called smart carts use sensors, in-store cameras, and computer vision software to keep track of which items get placed in or removed from carts, then automatically bill customers via their phones as they exit the store.
Analysts use AI-based software to adjust stock levels based on analyses of historical data, seasonality, promotions, and other factors.
Aisle-roaming robots read RFID tags on items to help retailers adhere to store planograms and fill empty spaces on shelves, even when workers initially misplace products. Store workers get notifications on their phones—complete with photos—of items that need to be moved or restocked.
AI systems help grocers design planograms, layouts of shelves, and floor displays based on analyses of shoppers’ previous routes through stores.
Food retailing, an industry long resistant to adopting emerging technologies, is under pressure to up its IT game in the face of increasing competition from big-box stores, discounters, and online specialists. The following challenges don’t make the industry’s move to AI any easier.
The grocery retail sector has been slow to adopt AI technology in part because of the costs. Installing cameras, sensors, and servers for computer vision powering automated checkouts can cost more than $1 million per store. Even short of that full-fledged automation, food retailers need access to considerable cloud computing resources to train and deploy AI models.
Training store associates to work on higher-value tasks associated with AI analyses is a considerable challenge. AI systems are in themselves complex and evolving rapidly, so training may have to be ongoing.
Grocery chains need to balance the collection of personal data about their customers and their buying habits with local and national rules that limit what businesses can collect, or else give consumers ways out of participating. Consumers in some markets need to explicitly opt in to receiving email or other electronic promotions, limiting what grocers can do. According to market research from Deloitte and Dutch supermarket chain Ahold Delhaize, 70% of consumers say they’re willing to share personal information with grocery chains. But the researchers recommend retailers don’t share that data further without explicit consent.
Oracle offers a range of software tools that can help grocers set inventory levels, shore up their supply chains, and show consumers dietary info and ingredients for private label brands.
Oracle Retail artificial intelligence and analytics can help grocery retailers optimize their inventory levels, identify merchandising opportunities, implement markdowns, and design marketing programs based on customer segments, demographics, and the performance of past promotions. Oracle Fusion Cloud ERP and SCM can help supermarkets forecast and manage their financials and strengthen their supply chains.
Oracle’s grocery sector applications include tools for making decisions on replenishing inventory and ensuring items stay in stock, sharing product information from store brands with shoppers, and implementing sustainability initiatives.
How is AI used in supermarkets?
Supermarkets use AI to improve the accuracy of their sales forecasts, get better performance from promotions and inventory placements, understand how sales of items influence those of others, tailor product assortments to local markets, make the best use of their store space, and perform other tasks.
How can grocery retailers use data to make better decisions with AI?
AI-based data analyses can help retailers avoid carrying too much inventory, stock products at stores where customers most desire them and restock them when they’re running low on shelves, target promotions more precisely, and set prices at levels that will maximize profits.
What are the benefits of grocery AI?
Grocery retailers ultimately benefit from AI by way of higher sales, fatter profit margins, and increased customer satisfaction.
How can AI help reduce food waste in supermarkets and grocery stores?
Some supermarkets use AI software to automatically lower the price of produce, dairy products, and other items that are near their expiration date or that aren’t selling, helping cut down on waste. They also use the technology to more accurately forecast demand.