From Narrative to Intent Schema Architecture: An Artificial Intelligence System for Human-Centered Advising Decision Support

Transfer advising is a high-uncertainty setting where students rarely ask well-structured questions and advisors must infer intent from fragmented narratives across conflicting institutional policies. This research presents a human-centered AI framework that transforms naturalistic student discourse into structured advising intents, supporting a student-facing advising assistant whose explainable outputs are designed to support advisor review. Using 19,668 transfer-related Reddit posts, it surfaces recurring advising intents via extraction, clustering, and structured labeling, then routes student questions to them. On a proof-of-concept benchmark it reached 90% intent-level and 86% question-level accuracy, turning advising knowledge into transparent structures that support scalable, human-centered advising.