Pilot study finds AI shopping trolley users spend 32% more in supermarkets

Smart shopping carts equipped with digital screens may be changing more than the way people navigate supermarket aisles. New research suggests that shoppers who use these connected trolleys spend substantially more, buy more items and remain in stores longer than customers who use conventional carts. In a month-long observational study involving more than 12,000 shopping sessions at a major German supermarket chain, smart-cart users spent an average of €30.18 per trip, compared with €22.89 among non-users. That difference amounts to approximately 32 percent, raising fresh questions about whether digital convenience is also becoming a powerful engine for impulse purchasing.

The carts, which are being trialled in major supermarkets, typically feature a tablet-like screen mounted on the handle. Shoppers can upload or create digital shopping lists, receive personalized product recommendations, obtain directions to particular aisles and, in some cases, complete payment without visiting a traditional checkout. The technology is designed to reduce friction: instead of searching for products, comparing routes or waiting in line, customers can be guided through a digitally assisted journey. Yet the same system can place promotions, recommendations and advertising directly in front of consumers at the exact moment they are making purchasing decisions.

Researchers from Bayes Business School at City St George’s, University of London, analyzed 12,418 individual shopping sessions recorded during March 2025. Of these, 9,422 sessions involved active smart-cart users, while 2,996 were completed by shoppers who did not use the technology. The researchers examined basket value, the number of products purchased and the amount of time spent in the store. Each session was also linked to the time of day and day of the week when the cart was activated, allowing the team to investigate whether the effects varied between mornings, afternoons, evenings, weekdays and weekends.

The results showed that smart-cart users bought an average of 12.02 items per trip, compared with 9.61 items purchased by non-users—a difference of roughly 25 percent. They also spent 40.40 minutes shopping, compared with 32.75 minutes among shoppers without digital assistance. The increases were especially pronounced during afternoon and evening trips, as well as on weekends. The pattern suggests that the technology may do more than help customers find what they already intended to buy. By extending the time shoppers spend inside the store and continuously presenting information, the carts may increase the number of opportunities for additional purchases.

From a behavioral science perspective, the finding is consistent with the idea that digital interfaces can alter the “choice architecture” of a physical environment. A shopping list normally limits attention to a set of planned products. A screen attached to the cart can expand that attention by presenting discounts, substitutes, complementary items and branded messages while a shopper is already moving through the store. Recommendations may be personalized using previous purchases, the contents of a digital list or the shopper’s location. Even when each individual prompt has only a small effect, repeated exposure can accumulate across a full shopping trip, increasing both basket size and total spending.

However, the study does not establish that smart carts directly caused shoppers to spend more. Because the research was observational, customers chose whether to use the carts rather than being randomly assigned to them. People who select a smart trolley may already differ from non-users in important ways: they may have larger households, greater interest in technology, more available time, stronger loyalty to the supermarket or a greater willingness to explore products. These pre-existing differences could contribute to the spending gap. The researchers’ findings therefore describe a strong association, while a controlled experiment would be needed to determine precisely how much of the increase is attributable to the cart itself.

The data also revealed a more complicated pattern among the most frequent users of the screens. Customers who interacted with the devices more than 20 times during a single trip—classified in the study as “superusers”—bought significantly more items and stayed in the store longer. Nevertheless, they did not spend more money overall. At very high levels of interaction, both spending and basket size began to decline. One possible explanation is that intensive users were engaging with the cart for entertainment, exploration or general screen activity rather than using it solely as a practical shopping tool. They may have examined more products, navigated more extensively or responded to numerous prompts without consistently adding expensive items to their baskets.

The researchers also observed environmental and temporal effects. Across all customers, higher temperatures were associated with shorter shopping trips, suggesting that weather can influence how long people remain inside supermarkets. Afternoon and weekend shoppers showed particularly strong spending differences between smart-cart users and non-users. The researchers argue that these periods may offer retailers especially valuable opportunities to promote premium products, new launches or targeted offers. Such strategies, however, also intensify the ethical debate surrounding digital retail: a system that makes shopping easier can simultaneously make it harder for consumers to recognize when convenience has become persuasion.

For retailers, the appeal of smart carts is clear. The devices can provide a direct communication channel inside the store, generate detailed information about customer interactions and support advertising that is more precisely timed than conventional aisle displays. They may also help stores reduce checkout congestion and connect physical shopping with online accounts and loyalty programs. For consumers, the technology can be genuinely useful, particularly for locating products, managing complex lists or completing a trip quickly. But the new findings suggest that convenience and commercial influence are becoming tightly intertwined. As supermarkets adopt increasingly intelligent interfaces, shoppers may need to treat digital recommendations with the same caution they apply to advertising elsewhere. The central question is no longer simply whether a smart cart can help someone shop, but whether it is also quietly encouraging them to shop more.

News Publication Date: 4-Jun-2026

Web References: Bayes Business School, City St George’s, University of London: https://www.bayes.citystgeorges.ac.uk/ ; Journal of Business Research article: https://www.sciencedirect.com/science/article/pii/S0148296326003723?via=ihub

References: Gottschalk, Sabrina; Lorenz-Kornfeld, Annika; Oc, Yusuf; and Eisler, Nils. “Customer responses to smart shopping carts in supermarkets.” Journal of Business Research.