More of the same: how recommendation cue redundancy in AI shopping assistants nudges consumer decision confidence

Purpose While AI shopping assistants aim to streamline consumer choices, they often create a paradox: abundant information can overwhelm consumers. This study investigates a counterintuitive digital nudging strategy – recommendation cue redundancy – which can bolster consumer confidence after an item has been added to a shopping cart. It further examines the underlying mechanisms and boundary conditions that shape this effect. Design/methodology/approach Four between-subjects experiments were conducted across diverse product categories. Statistical analyses were used to test the proposed mediation, moderation and sequential mediation effects. Findings High-redundancy (vs. low-redundancy) recommendation cues significantly increase consumer decision confidence. This relationship is fully mediated by information supportiveness. The effect is moderated by consumer risk tolerance: the positive impact is significantly stronger for consumers with low risk tolerance and is attenuated for those with high risk tolerance. This study also validates the effect of this strategy relative to the no-recommendation condition and demonstrates a sequential indirect effect on purchase completion intention. Originality/value This study extends digital nudging theory to the final decision stage of AI-assisted shopping by showing that recommendation cue redundancy can function as a stage-specific decision-assistance nudge. Practically, the findings offer guidance for designing AI shopping assistants that provide more effective and personalized support at the final stage of purchase.