Published Paper Details:

Higher education institutions (HEIs) handle a large volume of repetitive administrative tasks, such as responding to student queries, verifying academic records, scheduling, and document processing, that consume considerable staff time. Rule-based chatbots have partially automated such tasks, but they lack the multi-step reasoning and decision-making required for genuinely complex administrative workflows. This paper examines the application of Agentic Artificial Intelligence (Agentic AI), AI systems capable of planning, invoking external tools, taking autonomous actions, and verifying their own outputs, to the automation of academic administrative processes. We review the evolution of agentic architectures, from reasoning-and-acting paradigms to modern multi-agent orchestration frameworks, and propose a conceptual framework in which specialised agents handle distinct administrative functions (query resolution, records verification, scheduling, and escalation) under a coordinating orchestrator agent. We describe a proposed implementation approach using open-source agent-orchestration frameworks and synthetic academic data to avoid privacy risk, together with an evaluation methodology based on task-completion accuracy, response latency, and escalation correctness. The paper concludes with a discussion of feasibility, ethical considerations surrounding student data privacy, and directions for future empirical work. The proposed framework offers a low-cost, extensible pathway for resource-constrained HEIs, particularly in the Indian context, to reduce administrative overhead while retaining human oversight over sensitive decisions.