GenAI and Decision Making – An Eye-Tracking Study

With the rise of Generative AI (GenAI) in e-commerce, more and more platforms are integrating GenAI tools such as GenAI review summarization and chatbots to assist consumers in their decision making. Consumers experience less information overload and greater decision satisfaction when exposed to a manageable amount of user reviews (Hu et al., 2019). This suggests that GenAI review summarization can reduce the cognitive burden on a consumer. Research is now beginning to uncover how GenAI review summaries and their valence influence decision making (Wang et al., 2025; Todd et al., 2024). For example, Wang et al. (2025) found that AI-generated review summarization positively impacts product sales and that its effectiveness is moderated by the number of reviews (in a U-shaped pattern), average rating, and rating dispersion. These findings suggest that GenAI tools actively influence the purchasing decisions of a consumer under varying conditions. However, they examine such tools in isolation. The combined effect of more than one GenAI-powered tool in decision making, as in the real world, is yet to be explored. Our study investigates how GenAI review summaries and chatbot conversations collectively influence a consumer’s decision making. While GenAI review summaries give consumers a quick overview of what others are saying about a product or service, chatbots offer conversational responses tailored to a consumer’s specific questions. Drawing on Media Richness Theory, we theorize that chatbots represent a richer medium than static GenAI review summaries. We further study the definite valence of GenAI review summaries as a moderator of their influence on purchase intention. This study will use eye-tracking technology as a neurophysiological tool to measure a consumer’s cognitive engagement with each GenAI medium in an experimental setup. Participants will complete a shopping task on a simulated e-commerce website where each GenAI medium acts as a treatment. We will collect metrics based on predefined Areas of Interest (AOIs), such as fixation duration, fixation count, time to first fixation, etc. These metrics will be used to empirically analyze how consumers cognitively process each type of GenAI medium and their collective impact on decision making. Our findings will contribute to the growing GenAI and NeuroIS literature.