Voice assistant usage dilemma: balancing trust and privacy for user satisfaction among sandwich graduate students

Purpose This study examines the effect of voice assistant (VA) usage on user satisfaction, using trust and privacy as mediating mechanisms among graduate students in accelerated (sandwich) programs in Ghana. Design/methodology/approach Data were collected from 208 purposively sampled sandwich graduate students and analyzed using PLS-Structural Equation Modeling. Findings VA usage significantly influenced user satisfaction, trust, and privacy. Trust significantly affected user satisfaction, while privacy did not. Trust partially mediated the relationship between VA usage and satisfaction. Privacy did not mediate the relationship independently but did so jointly and sequentially with trust. Research limitations/implications The study extends technology acceptance and regulatory focus theories by demonstrating trust as a pivotal mediator in the VA usage–satisfaction link, with privacy exerting influence only through joint and serial pathways. This offers a nuanced understanding of trust–privacy mechanisms in emerging market contexts. Practical implications Universities should integrate VA tools into sandwich programs, while developers should embed both academic-relevant features and robust data protection mechanisms to enhance user trust and satisfaction. Originality/value As VA adoption among Ghanaian graduate students is still evolving, this study offers empirical insight into how trust and privacy influence satisfaction. It reveals that while students adopt VAs based on usefulness and ease of use, privacy concerns and prevention-focused mindsets may hinder satisfaction and continued usage.