Sales role play is a widely used method for training sales professionals, but it presents structural challenges related to time, convenience, reporting, and consistency. Traditional role play requires two individuals to coordinate schedules, assume buyer and seller roles, and may involve a third person for observation and feedback. Generative AI avatar sales training offers a potential solution to these challenges, yet research on best practices for Generative AI avatar sales training engagement and efficacy is limited or dated.

This study examines Generative AI avatar sales training through a mixed-method approach involving students and faculty in a large intercollegiate sales competition. Training engagement data from participants were analyzed in combination with competition results using statistical methods, while faculty provided qualitative insights through semi-structured interviews to contextualize the training environment. Sales faculty and intercollegiate sales students are like sales organizations in practice as sales faculty are de facto sales managers who often have significant practice experience and intercollegiate sales competition students are typically upper classmen who are destined for sales careers shortly.

Research findings reveal that, despite strong intrinsic and extrinsic motivation among students, just 40% of invited competitors utilized the offered training. Among those who did use the training, a statistically significant relationship between school experience, average role-play training session length, and competition performance was observed. Schools with the most experience in intercollegiate sales competitions had the greatest benefit from the training. The study considers Innovation Diffusion Theory, the Informing Science Rugged Landscape, and Self-Determination Theory to interpret these results and identify options to improve adoption of Generative AI avatar sales training for sales students and practice.