Algorithmic transparency is widely considered essential for fostering trust in AI-based financial e-commerce services. However, empirical evidence remains limited on whether transparency benefits all consumers uniformly and how it is evaluated relative to other service attributes in realistic decision contexts. This study examines how consumers trade off transparency, personalization, and user control in robo-advisor (RA) services across different consumer segments. Through a discrete choice experiment and latent class logit modeling, two distinct segments are identified: selective high-expertise investors, who prioritize personalization and user control over transparency, and receptive general consumers, who respond strongly to enhanced explainability. These findings indicate that algorithmic transparency does not serve as a universal design solution but operates conditionally based on consumer expertise and attribute interactions. Simulation results further show that while a regulation-compliant, uniform service design may facilitate market entry, it constraints long-term expansion in heterogeneous markets. In contrast, a segment-based service portfolio calibrated to the distinct preferences of each group significantly increases overall adoption under the same regulatory constraints. These results suggest that sustainable AI diffusion in financial e-commerce requires a nuanced approach that balances disclosure with functional autonomy to address the diverse needs of both sophisticated and novice users.
Choi et al. (2026) studied this question.