Abstract Purpose Perceived discomfort could indicate an early sign of pain, for example, as a result of a biomechanical load on the musculoskeletal system. Assessing discomfort can, therefore, help to identify workers at increased risk of musculoskeletal disorders for targeted intervention development. We aimed: (1) to identify the optimal cut-off value of neck and low back discomfort among office workers and (2) to evaluate its predictive validity with future neck and low back pain, respectively. Methods At baseline healthy participants (n = 100) completed questionnaires, including the Borg CR-10 discomfort scale (on a 0–10 scale), and were followed for six months, during which musculoskeletal pain was assessed monthly. Logistic regression analyses were performed to assess the associations of baseline discomfort with the onset of future neck or low back pain. Sensitivity, specificity, and the area under the receiver operating characteristics curve were estimated to identify the optimal discomfort cut-off value predicting future pain. Results Borg CR-10 scores ≥ 3.5 for perceived neck and low back discomfort had acceptable sensitivity and specificity to predict future neck and low back pain, respectively. Perceived discomfort at baseline as a dichotomous measure (using the ≥ 3.5 cut-off) was a statistically significant predictor of future neck pain (OR = 10.33) and low back pain (OR = 11.81). Conclusion We identified the optimal cut-off value of the Borg CR-10 discomfort scale to identify office workers at increased risk of developing neck and low back pain. These findings might benefit ergonomists, primary health care providers, and occupational health researchers in developing targeted interventions.