<p>Effective coach-athlete feedback is essential for skill development, yet most automated coaching systems still fall short of the nuanced personalization and context-awareness a human coach provides. We introduce an intent-aware personalized feedback generation framework that fuses sports-education insights with modern NLP methods. The system first categorizes each coach utterance into a communicative intent (e.g., tactical instruction, encouragement, correction) and then produces a feedback message tailored to the athlete’s individual profile. By guiding generation with explicit intent recognition, our approach ensures that both the tone and content of the output faithfully reflect the coach’s communicative goals. We formalize the task through a two-module architecture that pairs an intent classifier with a persona-based neural generator. Evaluations on real and simulated coach-athlete dialogues show that the proposed model yields feedback that is consistently more relevant and personalized than strong baselines, according to automatic metrics, expert judgments, and practitioner user studies. Ablation studies further highlight the distinct contributions of intent modeling and athlete profiling, while qualitative analyses confirm the feedback’s practicality and motivational value-suggesting meaningful benefits for sports training.</p>