This article describes the development of a tutoring support system (TSS) for online teacher training courses in Uruguay. The developed system interoperates with the learning management system (LMS) and automates various tasks, in order to optimize the time and effort of tutors involved in online teacher training courses. Among its features, the AI-assisted pre-correction and feedback module stands out. This module automatically pre-corrects assignments completed by participants in each course, which tutors then review and validate before final delivery. This paper analyzes the operation of this module using data collected during 2024 and 2025. It is observed that improvements in prompts, as well as advancements in the models, lead to increased efficiency of the AI-based module.