In AI We Trust: Characteristics Influencing Assortment Planners’ Perceptions of AI Based Recommendation Agents

Abstract

While creating an optimal assortment of products, assortment planners need to take into account an important amount of information, which leads to a certain level of uncertainty. These trade-offs can diminish the quality of the assortment decisions made by the planners. To reduce their impact, assortment planners can now use artificial intelligence (AI) based recommendation agents (RAs) throughout their decision-making process, thus benefiting from their ability to process a large quantity of information to improve their decisions. However, research on user-RA shows that there are some challenges to their adoption. For instance, RA adoption depends on the users perceived credibility of its recommendations. Hence, this study investigates how the richness of the information provided by the RA and the necessary effort to access this information influences the assortment planners’ usage behavior (visual attention) and perceptions (credibility, satisfaction, performance, intention to adopt the RA). A within-subject lab experiment was conducted with twenty participants. The results show the importance of the RA’s recommendations that include easily accessible explanations of the variables included in their calculations on the usage behavior, perceptions, and decision quality of the assortment planners. These findings contribute to the HCI literature and the theory of RA adoption in B2B contexts by providing insights on features enhancing employee adoption.