Agentic AI Workflows for Teaching and Learning, Research, and Community Service: A Cross-Disciplinary Framework for Tertiary Education Adoption

Abstract

The emergence of agentic artificial intelligence (AI) systems capable of autonomous goal pursuit through planning, tool use, persistent memory, and multi-step execution represents the most profound transformation of institutional practice since the advent of the internet. Unlike prior generations of AI tools that respond to discrete prompts, agentic systems can initiate, coordinate, and complete entire workflows with minimal human intervention. For the modern tertiary education institution, which fulfils simultaneous obligations in teaching and learning, knowledge production, and community service, this transformation is not merely a question of efficiency but of fundamental capability expansion. This paper presents a comprehensive, cross-disciplinary scholarly analysis of agentic AI deployment across all three institutional functions. Drawing on verified, peer-reviewed literature from 2022 through 2025, the paper examines: (1) agentic workflows in Teaching and Learning, including instructional content generation, AI-assisted essay marking using rubrics, and adaptive pedagogical applications; (2) agentic AI in Research, with emphasis on literature synthesis, Zotero integration via the Model Context Protocol (MCP), research proposal pipelines, and citation verification; and (3) Community Service applications encompassing neighbourhood security and home automation. A dedicated section addresses ethical risks — including academic integrity threats, algorithmic bias, surveillance overreach, and equity of access — and proposes a governance framework grounded in institutional best practices. The paper concludes that agentic AI, deployed responsibly within a framework of institutional governance, AI literacy, and human oversight, represents an extraordinary opportunity to fulfil the university's three-fold mission with unprecedented scale and quality.