This study investigates the use of artificial intelligence (AI) to support written feedback in English for Specific Purposes (ESP) in higher education through a pedagogically guided automated system developed with n8n and AI tools. Implemented in a B2-level university course, the system integrated a teacher-designed rubric aligned with CEFR criteria to provide immediate, structured feedback on students’ written texts. Results show a high degree of alignment between AI-generated feedback and teachers’ assessments, particularly in grammar, vocabulary, and textual coherence. Students reported positive perceptions of the feedback’s clarity, usefulness, and motivational impact, while teachers observed a significant reduction in correction time without loss of formative quality. The experience also encouraged critical reflection on the ethical and pedagogical implications of AI in assessment. The study demonstrates that AI, when embedded within a pedagogically controlled framework, can enhance formative assessment, support student autonomy, and complement rather than replace teacher judgment, offering a transferable model for higher education.