BACKGROUND: Graduate medical education is a critical period for fostering self-directed learning (SDL). This study introduced an academic coaching program to support SDL among internal medicine (IM) residents, leveraging Gallimore and Tharp's four-stage model as a scaffolding framework. OBJECTIVE: To assess the impact of academic coaching on residents' performance, including Internal Medicine In-Training Examination (IM-ITE) scores, individualized learning plans (ILPs), and attitudinal changes. The study also explored how coaching influenced SDL within the residency program. DESIGN AND PARTICIPANTS: A mixed-methods case study was conducted in a mid-sized university's IM residency program. Quantitative measures included pre- and post-coaching surveys, ILP analysis, and IM-ITE score evaluation. Semi-structured interviews provided qualitative insights into participant experiences. Of 77 eligible residents, 40 enrolled in the coaching program, and 27 (18 post-graduate year (PGY) 1 and PGY2) completed at least one session. Baseline IM-ITE scores guided enrollment for mandatory participants. KEY RESULTS: Of the 77 residents, 51 had complete IM-ITE data and individualized learning plans (ILPs) from 2022 and 2023. Residents attending one coaching session demonstrated significant improvement in IM-ITE percentile scores (p = .022), while those with two or more sessions showed significant gains in both percent correct (p = .015) and percentile scores (p = .003). No significant differences were observed in ILPs or attitudinal surveys. Qualitative analyses of resident participant interviews highlight coaching's positive influence on SDL, organized into input, process, and output domains. CONCLUSIONS: Sustained coaching, defined as two or more coaching meetings, is associated with improved IM-ITE performance. Qualitative findings underscore the program's role in enhancing residents' SDL.