Applying an Omni-Channel Agentic AI Model to Optimise University Admissions Counselling and Data Capture

Abstract

This paper proposes an omnichannel agentic AI model to improve admissions counselling and lead data capture in university recruitment, where prospective applicants increasingly seek information across fragmented digital platforms. Unlike conventional, script-driven chatbots, an LLM-based agent can reason, plan a multi-turn dialogue, and invoke tools to: (i) provide accurate responses grounded in an institution-approved knowledge base; (ii) proactively standardise and capture contact details; and (iii) trigger handover to admissions officers through a human-in-the-loop (HITL) mechanism. The paper presents a three-layer reference architecture (omni-channel interaction—agentic cognition and orchestration—data and action), an operating workflow, and an illustrative implementation configuration aligned with the resource constraints of Vietnamese universities. The study concludes by discussing the managerial implications and requirements for safety, accountability, and personal data protection.