Entrepreneurial passion can influence individual well-being and improve firm-level outcomes, yet little is known about how to rapidly detect a change in passion from entrepreneurs’ communication. We draw on advancements in both the passion literature and artificial intelligence (AI) methods, to capture entrepreneurial passion expressed for founding a venture at different points in time. Specifically, we developed an AI algorithm to recognize identity-based passion (identity centrality) from training data, comprised of eight hours of transcribed interviews with entrepreneurs (achieving 84% accuracy), and detect affective passion (intense positive feelings) with sentiment analysis. Application of these two novel measurement approaches, to longitudinal interview text with early-stage entrepreneurs (N=11, two time periods) in a six-month social venture accelerator, indicate that intense positive feelings decline while identity centrality varies. We conclude by outlining opportunities for future research.