Traditional score-based assessments often fail to capture the full efforts of student engagement, especially in hands-on courses. This work introduces a framework using Agentic AI to create a more supportive and insightful evaluation system. It emphasizes how AI can enable more effective and personalized evaluations in hands-on learning. By analyzing a student’s entire “learning portfolio”—including projects, formative assessments, and other artifacts—our approach provides personalized feedback, reduces test anxiety, and helps teachers understand individual learning processes. Our system features an AI agent integrated with a Learning Management System (LMS). This agent autonomously collects and analyzes student work, using a Large Language Model to synthesize the data into simple reports. This automates grading and frees instructors to focus on teaching higher-order skills rather than administrative tasks. While challenges like potential AI bias, data privacy, and over-reliance on technology must be addressed, this approach represents a promising shift in education. By moving from a single final score to assessing the entire learning process, we can create a more comprehensive “learning value” for each student. This approach better prepares students for real-world challenges and fulfills the assessment requirement of developing a more precise, meaningful, and supportive system for the hands-on learning curriculum.