The purpose of the current study was to determine whether AI can assist sport managers in the challenging space of salesperson training through the evaluation of interactions with prospective ticket buyers.
A total of 397 inside sales calls for a National Basketball Association (NBA) team were transcribed and analyzed. Unsupervised topic modeling was conducted using Latent Dirichlet Allocation (LDA), a machine learning technique for identifying latent themes or topics within a textual corpus. Additionally, sentiment analysis was conducted using text mining to generate sentiment polarity scores for sales call interactions. Finally, panel analysis through Ridge regression was conducted to assess appropriate model fit across evaluations of sales interactions.
Findings showed that sales agents who use a more diverse vocabulary are more likely to generate positive customer responses and achieve successful sales. Sales representatives who employ a more diverse vocabulary (higher lexical diversity) tend to exhibit positive sentiment polarity. Additionally, a moderate speaking pace contributed to more favorable customer perceptions. Finally, a positive association between the number of questions asked per call and open-ended questions was found, indicating agents who engage in more in-depth questioning strategies are better equipped to gather comprehensive information from customers.
Guided by the Technology-Task-Fit theory, the current study examined the role of AI-analysis in assessing sport ticket sales calls and assisting in salesperson training. The current study provides support for the use of AI in sales training through analyzing sales conversations and detecting differences in sales calls among successful and unsuccessful outcomes. These results demonstrate that AI can effectively analyze sales calls and provide evidence-based applicable feedback to sales agents.
