An On-Demand Solution for Scalable Reflective Tutoring Using Customised AI Agents
Date
Authors
Chanane, Nawal
Kuo, Matthew
Supervisor
Item type
Conference Contribution
Degree name
Journal Title
Journal ISSN
Volume Title
Publisher
Australasian Association for Engineering Education (AAEE)
Abstract
CONTEXT: Generative Artificial Intelligence (AI) has been increasingly explored in engineering education to support student learning and reflection. In a second-year Software Development Practice course at Auckland University of Technology (AUT), Auckland, New Zealand, we developed and deployed an AI agent called the Upskilling Log guidance agent using the Cogniti platform, to help students develop and refine their Individual Upskilling Logs. Students submitted individual upskilling logs with reflective content as part of their Sprint 0 project preparation, allowing educators to focus on more in-depth mentoring during laboratory sessions. PURPOSE OR GOAL: The purpose of developing the Upskilling Log guidance agent was to address students' uncertainty about structuring their logs and effectively incorporating technical content. The agent's prompts were specifically designed to provide targeted feedback and guidance that supported independent critical thinking, self-assessment, and reflective writing skills, while maintaining academic integrity and ensuring alignment with project deliverables. APPROACH OR METHODOLOGY/METHODS: This paper uses a design-based research approach to document the iterative development of an AI agent. The team, consisting of the course lead and a senior learning designer, tested and refined the agent over 12 weeks without student data. The study focused on designing effective prompts and analysing responses across iterations. The Upskilling Log aimed to balance technical and reflective writing, encouraging students to critically evaluate their learning independently outside class. ACTUAL OR ANTICIPATED OUTCOMES: Through iterative development and testing, the Upskilling Log agent was successfully refined to help students structure their reflections around five key areas: development environment learning, team collaboration dynamics, Sprint 1 User Story readiness, areas for continued development, and insights gained during upskilling. Based on the researchers' observations and prompt reviews, students who engaged meaningfully with the agent demonstrated improved log structure, enhanced alignment with project requirements, and reduced dependence on educators for basic feedback guidance. CONCLUSIONS/RECOMMENDATIONS/SUMMARY: This development showed AI tools like Cogniti can effectively design agents that enhance reflective practice and self-directed learning in engineering education through systematic, iterative refinement and clear pedagogical frameworks. It highlighted the importance of careful prompt design, defined boundaries, and balancing AI support with educator guidance. This researcher-led approach offers a transferable model for creating AI agents that foster critical reflection while maintaining academic integrity.Description
Keywords
Generative AI, Reflective Practice, Engineering Education, Self-Directed Learning
Source
The 36th Australasian Association for Engineering Education Annual Conference. 7 - 10 December 2025 | The University of Queensland, Brisbane
DOI
Publisher's version
Rights statement
Copyright © 2025 Nawal Chanane, Matthew Kuo: The authors assign to the Australasian Association for Engineering Education (AAEE) and educational non-profit institutions a non-exclusive licence to use this document for personal use and in courses of instruction provided that the article is used in full and this copyright statement is reproduced. The authors also grant a non-exclusive licence to AAEE to publish this document in full on the World Wide Web (prime sites and mirrors), on Memory Sticks, and in printed form within the AAEE 2025 proceedings. Any other usage is prohibited without the express permission of the authors.