This paper quantifies labor-related exposure at the firm level by directly extracting labor discussions from earnings conference call transcripts using OpenAI GPT-4o model. The measure captures the relative emphasis on labor issues during management-analyst interactions, reflecting both management’s priorities and analysts’ concerns. It tracks macroeconomic trends, distinguishes between labor- and capital-intensive industries, and is validated against multiple firm labor-related outcomes. We find that firms with higher labor-related exposure scores exhibit lower labor productivity, lower market-perceived labor efficiency, slower subsequent employment growth, and higher likelihood of labor-related misconduct, as well as more severe labor-related penalties. Labor-related exposure is also positively associated with realized stock return volatility, confirming its relevance as a source of firm-specific uncertainty. Our analysis of corporate outcomes shows that higher labor-related exposure correlates with poorer future operating performance, as evidenced by lower return on assets (ROA), lower return on sales (ROS), and lower operating cash flows. We also identify a dual role of labor-related exposure in capital allocation: firms with higher exposure scores tend to reduce capital investments while simultaneously increasing R&D spending in the near term. This suggests that heightened labor-related exposure motivates firms to reallocate resources toward efficiency-enhancing innovations while suspending less attractive projects. Our findings highlight labor-related exposure as a distinct dimension of firm-level uncertainty and demonstrate the potential of large language models (LLMs) in advancing empirical research on labor dynamics.