Empathy is a key predictor of organizational and individual success, and recent advances in machine learning enable the automatic and efficient extraction of such social signals from publicly available digital data. Emotional mimicry, defined as the rapid, often unconscious alignment of one person’s expressed affect (e.g., facial emotion) with another’s, has been identified as a key facilitator of empathy that significantly influences interpersonal relationships and leadership effectiveness. Building on neuroscience research, we introduce a video analytics framework that estimates an Affective Mimicry Index (AMI) as a behavioral proxy for empathy. To illustrate the effectiveness and practical value of AMI in a real-world setting, we analyze televised CEO interviews in which managers address questions about business success and performance. For each CEO, we construct the AMI by quantifying emotional mimicry toward the interviewer within each question–answer pair and aggregating across the interview. We then examine how this video-based empathy proxy for CEOs is associated with corporate outcomes. Our findings reveal that CEO AMI is positively related to workplace safety. Additionally, firms led by CEOs with higher AMI exhibit higher firm value and fewer regulatory penalties for misconduct. These results suggest that more empathic CEOs are more likely to enhance employee welfare, prevent organizational crisis, and increase firm value. Methodologically, this paper contributes to business analytics and FinTech by integrating large language models, conversational analytics, and computer vision to capture empathy from video recordings. The theoretical and managerial implications of our study are discussed.