As artificial intelligence continues to rapidly reshape the world around us, including the grocery shopping experience, a new research study from the University of Florida Department of Food and Resource Economics (FRE), reveals human touch may be a crucial component for food retailers trying to retain their customer base.

While the vast majority of major food retailers already utilize some form of partially automated self-checkout system, several major businesses such as Amazon and Sam’s Club have started to introduce a fully automated experience.

“When it first entered the market, it generated considerable excitement and debate within the grocery retail industry,” Chen said. “Some people worried that this format would threaten cashiers’ jobs, while others viewed it as a sign of technological progress in grocery retail.”

The study, published in the Journal of Retailing and Consumer Services, was led by Junhong Chen and her advisor, UF Institute of Food and Agricultural Sciences (UF/IFAS) FRE Professor Zhifeng Gao.  The research team used a nationwide survey conducted in June 2021 to determine whether attributes like local or organic food availability, full or partial automation and social responsibility would impact how far a customer was willing to travel to the grocery store.

“This topic is particularly timely given the evolving reception of Amazon Go, a fully automated grocery store format,” said Junhong Chen, lead author, whose work on the study was completed at UF’s Institute of Food and Agricultural Science during her Ph.D. program at FRE. “Over time, a growing number of Amazon Go locations have closed, raising questions about the long-term viability of this retail model.”

Results revealed that a significant factor impacting consumer store preference was the desire to avoid a fully automated store format.

While consumers did prefer a partially automated experience where self-checkout was an available option, they were much less willing to shop at a store with no cashiers at all. Results showed a 21-minute reduction in travel time would be needed to convince participants to use the fully automated grocer.

This preference was shown across all demographics represented in the study, suggesting a consensus across consumers.

“The strong preference for a partially automated store suggests that consumers seek a balance,” Chen said. “They want the convenience of automation but also the reassurance of readily available human assistance when service needs arise.”

AI-enabled technology in grocery stores is new, and some of the consumer hesitancy may be traced to the lack of familiarity and the limited information available. To test this, the researchers presented different survey respondents with either positive or negative information about this type of shopping experience.

While negative information seemed to have no effect on consumer perception of fully automated stores, the impact of positive messaging differed based upon whether the benefits were perceived as belonging to the individual shopper or broader society.

When presented with messaging emphasizing the societal benefits that store automation may create new employment opportunities in technical fields, consumers had a higher willingness to travel to these stores.

However, when survey participants were given information about the potential positive personal impacts such as faster shopping, the absence of checkout lines, and seamless transaction completion via store apps, the willingness to travel decreased.

This aligns with a broader emphasis seen on store-level ethical practices in determining consumer shopping preferences seen elsewhere throughout the study, as among all attributes tested in the survey, store ethical ratings exhibit the strongest positive effects on store choice.

“For food retailers, this means that decisions about adopting new technologies must account for not only potential cost savings but also consumer resistance to innovations perceived as too radical,” Chen said.

This work was supported by the Florida Agricultural Experiment Station and the USDA-NIFA (Hatch/multistate project FLA-FRE-006726).