Empowering Senior AEC Students: A Real-Time, LLM-Driven Tool for Technical Writing Feedback

Technical writing is a major challenge for senior AEC students, who must convey complex concepts accurately. Most writing support focuses on foundational courses, leaving students underprepared for capstone projects and prone to overreliance on unsupervised LLMs. This study introduces an LLM-driven feedback tool that offers immediate, rubric-based guidance—emphasizing correctness, clarity, completeness, and consistency—while preventing the tool from generating content. Deployed in a senior Capstone Project course via a Flask-based client–server model, the system supports both chat submissions and file uploads. A comparison with human TA feedback revealed similar detection of grammatical and structural issues, with TAs adding domain-specific insights. Results indicate that combining automated, iterative feedback with TA expertise can improve writing quality, foster critical thinking, and help students meet professional standards.