Agentic AI in Higher Education: Developing a Custom Classroom Engagement Platform Through AI-Assisted Software Engineering

Abstract Book of the 11th International Conference on Advanced Research in Teaching and Education

Year: 2026

[PDF]

Agentic AI in Higher Education: Developing a Custom Classroom Engagement Platform Through AI-Assisted Software Engineering

Prof. Dr. Arkhe Pacis, Jacob Bramwell

ABSTRACT:

Institutional learning management systems (LMS) often fail to adapt to the granular, instructor-specific workflows that animate active classrooms in business analytics and operations management courses. Recent advances in agentic AI — autonomous, goal-directed systems built on large language models that can plan, reason, and write code with minimal human oversight — are reshaping how software is conceived and built. This paper presents the design, development, and operational deployment of PacisClassPass, an AI-enhanced classroom engagement platform built by an instructor for undergraduate business analytics and operations management courses. The platform combines AI-generated quiz items, mobile push notifications, real-time classroom synchronization, GPS-verified attendance with selfie capture, reflection submissions, and instructor-facing analytics dashboards, deployed as a progressive web application (PWA) over a Node.js, WebSocket, MongoDB, Redis, and Firebase Cloud Messaging stack. Methodologically, the project follows a Design Science Research (DSR) frame: the platform is the artifact, and its iterative construction via agentic AI workflows is the contribution. We document the agentic AI development workflow — iterative prompting, LLM-assisted code generation, debugging collaboration, deployment refinement — and report deployment observations, implementation challenges, and operational considerations across a full semester of instructional integration. The paper does not make causal claims about student learning outcomes; instead it offers implementation insights into how faculty, working as citizen developers augmented by agentic AI, can produce custom educational infrastructure that complements rather than replaces institutional LMS systems. We close with implications for the democratization of educational software development and a research agenda spanning adaptive learning, predictive analytics, LMS interoperability, and intelligent classroom orchestration.

Keywords: Design Science Research; Educational Technology; Human–AI Collaboration; Mobile Learning; Progressive Web Applications