This study investigates the development and implementation of Prompt Mentor, an AI-powered feedforward tool designed to support students in developing effective prompts in using and engaging critically with generative AI. The tool provides structured, iterative guidance to enhance academic reasoning and writing, and is integrated into an engineering materials course via RMIT’s institutional chatbot (VAL). Prompt Mentor was developed in alignment with the CDIO educational framework and ethical standards for secure, scalable deployment within the university’s learning environment. Two research questions guided the study: (1) how can Prompt Mentor be integrated into an engineering course to support prompt formulation; and (2) what pedagogical principles enhance reasoning and critical thinking through using such tools? Analysis focused on the quality of student-generated prompts, the use of AI feedback, and evidence of critical engagement in submitted assignments and reflections. Findings indicate that Prompt Mentor supported the development of prompt engineering as a teachable, iterative skill. Students demonstrated an improvement in their ability to frame specific and relevant inquiries and apply AI responses thoughtfully in their academic work. The tool functioned effectively as a feedforward mechanism, encouraging structured inquiry, self-regulated learning and increased confidence, particularly among students from non-English-speaking backgrounds. This study contributes new insights into the pedagogical design and institutional integration of generative AI in higher education. It moves beyond a technical demonstration to provide evidence of learning outcomes and transferable design principles for AI-supported academic development. Our study confirms that Prompt Mentor offers a replicable model for ethically grounded, pedagogically aligned AI tools that can address scalability and inclusivity in engineering education.
