Evaluating a generative AI-based pitch simulation tool: a pilot study in entrepreneurial training

Entrepreneurship education faces persistent challenges in providing scalable, standardized, and feedback-rich training for investor pitch skills. This pilot study developed and evaluated a generative AI-powered pitch simulator designed to replicate investor interactions in a low-pressure training environment.

A mixed-methods feasibility evaluation was conducted with 46 entrepreneurs. The study integrated post-session quantitative measures of perceived utility and learning value, qualitative thematic analysis of open-ended responses, and behavioral engagement data (interaction duration and message frequency).

Participants reported moderate perceived learning value (M = 6.02, SD = 2.30) and helpfulness (M = 6.39, SD = 2.14), with moderate-to-positive willingness to recommend the tool (M = 7.00, SD = 2.53). Qualitative analysis identified the simulator's value as a repeatable, structured practice environment, while a subset of participants also noted reduced emotional and nonverbal fidelity compared to face-to-face investor meetings.

The simulator functions as a scalable supplementary tool for entrepreneurship education, enabling iterative practice in a standardized setting. The findings also suggest that combining simulator practice with human mentorship may be particularly valuable for addressing interpersonal dimensions of pitch performance.

This study contributes an initial empirical evaluation of a generative AI pitch simulator and offers preliminary observations on how such tools might support experiential learning approaches in entrepreneurship education, providing a basis for more rigorous future investigation.