BuddyGPT

BuddyGPT is an innovation project (funded by BMS-WSV innovation research fund 2024) that connects conversational AI with Learning Management Systems (such as Canvas) to provide quick and accurate answers about courses. Within the project, a wide range of features has been designed to ground chatbot functionality in educational theories, thereby supporting effective mechanisms underlying learning processes.

The first prototype of BuddyGPT has been successfully piloted within a joint module at the University of Twente (M3 – Business Process Analytics) across two educational programmes - Industrial Engineering and Management (B-IEM) and Business Information Technology (B-BIT), and tested with over 150 students participating in the evaluation, additionally including students from Psychology program.

BuddyJ is a lightweight, private companion to BuddyGPT, designed for local deployment on a user’s laptop. It enables intelligent question-answering over a curated, smaller set of resources, making it ideal for personal, course-level, or project-specific use. By operating locally, BuddyJ supports data privacy, fast access (cloud model), and offline capabilities (local model), while retaining the core feedback-driven philosophy of the BuddyGPT ecosystem.

With support from BMSLab, a testable demo version of the Course Assistant is available for exploration. This version uses the local private companion BuddyJ, which is a pre-configured to the M3 – Business Process Analytics module. For privacy and compliance reasons, the demo is connected to a limited, curated set of resources within the module, including the module homepage, module manual, and selected course pages. It does not have access to course files or administrative pages through which additional courses or resources can be configured. Access to the demo requires being on the UT network or connected via VPN.

To configure BuddyGPT for your own course, or to participate in testing that informs ongoing research, please contact Gayane Sedrakyan.

Thesis topics related to the project are available, and research collaborations are warmly welcomed.

Theoretical background 

[1]    Sedrakyan, G., Malmberg, J., Verbert, K., Järvelä, S., & Kirschner, P. A. (2020). Linking learning behavior analytics and learning science concepts: Designing a learning analytics dashboard for feedback to support learning regulation. Computers in Human Behavior, 107, 105512.

[2]    Sedrakyan, G., De Weerdt, J., & Snoeck, M. (2016). Process-mining enabled feedback:“tell me what I did wrong” vs.“tell me how to do it right”. Computers in human behavior, 57, 352-376.

Publications from the BuddyGPT project: 

[3]    Design Implications for Next Generation Chatbots with Education 5.0 (2024)
In New Technology in Education and Training: Select Proceedings of the 5th International Conference on Advance in Education and Information Technology (pp. 1-12) (Lecture Notes in Educational Technology; Vol. Part F3326). Springer. Sedrakyan, G., Borsci, S., van den Berg, S. M., van Hillegersberg, J. & Veldkamp, B. P.  https://doi.org/10.1007/978-981-97-3883-0_1

[4]    Design Implications for Integrating AI Chatbot Technology with Learning Management Systems: A Study-based Analysis on Perceived Benefits and Challenges in Higher Education (2024) In ICAITE 2024: Proceedings of the 2024 International Conference on Artificial Intelligence and Teacher Education (pp. 1-8). ACM Press. Sedrakyan, G., Borsci, S., Machado, M., Rogetzer, P. & Mes, M.  https://doi.org/10.1145/3702386.3702405

[5]    Designing Explainability Features for LLM-based Educational Chatbots to Promote Reflective Learning Behavior (2025) [accepted]. XAI-Ed 2025: Pedagogy-Founded Explainable AI for Transparent, User-Centered AI in Education, 26th International Conference on Artificial Intelligence in Education, Costea, I. & Sedrakyan, G.

[6]    Usability testing of BuddyGPT : A proprietary and integrated conversational agent to support the learning of university students -  Li, Jiawei (Thesis, 2025) 

PhD thesis inspired by BuddyGPT project and instrument by a visiting PhD student:

[7]    N. Baz Aktas (visiting PhD researcher – promoted): Meta-Requirements for Enhancing User Experience in Conversational Agents: A Design Science Research Approach (2025)

Invited Talks on BuddyGPT (National and UT cross-faculty events):

[8]    Didactic use of AI in Statistics Education 

[9]    DSI Meet-up: AI Chatbots in Education - Lessons from the BuddyGPT Project 

[10] Poster presented at AI in Education Walk-in session (Faculty of Science and Technology)

Pilots and demos with support of BDSI / BMSLab server

[11] BuddyGPT Project and tool description: BuddyGPT | Living Models Lab

[12] BuddyGPT demo: BuddyGPT How to chat with the chatbot - YouTube

[13] BuddyGPT spinoff (BuddyJ locally deployable version with limited resources): Course Assistant