Dr. Ramon Lawrence, Professor, Department of Computer Science, Mathematics, Physics, and Statistics, Faculty of Science, UBC Okanagan
Courses: COSC 304, Introduction to Database Systems, INDG 100, Introduction to Decolonization: Indigenous Studies, and others
Our group developed HelpMe, an AI-enhanced student support platform that unifies multiple help channels including an AI chatbot, asynchronous questions, and office hours into a single system. The tool was created to address the challenge of supporting (large) classes with limited instructional staff while ensuring that students receive timely, course-specific guidance. By embedding instructor validation into everyday workflows, HelpMe keeps AI answers accurate, trustworthy, and aligned with course expectations. Students have responded positively, valuing the system as a central place to seek help, while instructors have noted fewer repetitive email questions and the growth of a reusable knowledge base that strengthens the AI over time. The system has been deployed in courses such as COSC 304, INDG 100, and some education courses, with broader rollout planned for 2025. Since HelpMe is developed and hosted at UBC, it ensures privacy and institutional control while also enabling unique research into how students and instructors can most effectively use AI in education.
Faculty interested in trying HelpMe can go to coursehelp.ubc.ca.
What is the AI tool for student support?
HelpMe is an integrated student support system that combines AI with existing help options such as office hours and TA support. It centralizes all help requests into one platform, ensuring students have 24/7 access while instructors and TAs remain in control of validating AI responses.
What motivated the development of this tool?
Large computer science courses (often 200+ students) generate hundreds of student help requests. Email and office hours alone could not scale without increasing staff workload. Commercial AI assistants were available but often gave generic or unreliable answers. HelpMe was created to reduce repetitive questions, improve accessibility of help, and align AI responses with course-specific content under instructor oversight.
What opportunities or challenges have emerged with the development of this tool?
Opportunities:
- Providing students with on-demand support, including evenings and weekends.
- Enabling instructors to transform validated answers into reusable knowledge, improving AI accuracy over time.
- Analyzing insights from student interactions provides information on learning difficulties and helps understand how AI can be effectively used to support student learning.
Challenges:
- Ensuring transparency so students and instructors trust the AI.
- Balancing AI speed with instructor oversight (AI answers may need to be checked or validated).
- Addressing ethical and privacy issues when logging student questions.
What are the benefits of this tool for students and instructors?
Student surveys indicated strong appreciation for having a single, unified place to ask questions, with the system rated highly for effectiveness. The AI component provided round-the-clock answers that were tailored to the course, giving students timely and relevant support.
For instructors and TAs, validated AI answers reduced repetitive work and created a scalable knowledge base. Importantly, integrating validated Q&A pairs increased factual correctness of chatbot answers by 24.5% compared to course-content only. These validated answers are recorded automatically as instructional staff help answer questions for students.
How have students responded to this tool?
Student response has been positive. The majority of the class used the help queue feature while about 50% used the AI features. While students still preferred human help for complex or personal queries, they valued the AI for quick clarifications. Survey comments noted that the platform made it easier to find help without having to guess the right channel.
How does this tool align with student support policies?
HelpMe is designed to support instructors, not replace them, by reducing repetitive questions and streamlining student support. The research was approved through UBC’s ethics process and reviewed under a privacy impact assessment. This ensures that HelpMe enhances student support while aligning with UBC’s commitments to accessibility, equity, and the ethical use of technology.
What advice would you give to other educators interested in developing similar tools to support student learning or teaching?
The best advice is to start small and build up over time. The HelpMe system began as a single-course pilot and expanded as student and instructor feedback shaped its development. It is important that any system be convenient and easy to use, and integrating with existing platforms (such as Canvas and CWL) helps adoption. Most importantly, keep instructors “in the loop.” AI in education is most effective when it continually learns from instructor input and feedback. Even small contributions from faculty can compound over time, leading to more accurate AI support and reduced instructional workload.