All Projects
Production

Student Services Assistant — Human-in-the-Loop AI

Secure AI-assisted SPA converting free-text student requests into structured, staff-reviewed service tickets — responsible AI with humans in the loop

JavaScript SPA AI/NLP Natural Language Processing Human-in-the-loop AI REST API Authentication Role-based Access

Overview

Built a secure, AI-assisted single-page application for universities that converts unstructured free-text student requests into structured, actionable service tickets reviewed by staff before any action is taken. The system is designed on the principle that AI should augment human judgment, not replace it — especially in high-stakes institutional contexts.

The Challenge

University service desks deal with thousands of unstructured student requests: financial aid questions, housing disputes, transcript requests, registration issues. Staff spend significant time deciphering what students actually need before they can help. Fully automated chatbots lack accountability and miss nuance. The challenge was to capture AI's language understanding capability while keeping the human firmly in the decision loop.

Technical Approach

  • AI-powered NLP layer to parse free-text student requests into structured categories and sub-categories
  • Ticket generation: extracts entity information (student ID, course, issue type) automatically
  • Staff review queue: all AI-generated tickets require human review before action
  • Human-in-the-loop design: AI handles classification and structuring, staff handle decisions
  • Secure SPA architecture: authentication, role-based access, audit logging
  • Designed for operational efficiency: faster intake, clearer requests, better staff focus on solutions
  • No autonomous chatbot responses — no AI-generated answers sent to students directly

Key Outcomes

Human-in-loop Responsible AI
Structured From Free-Text
Staff Review Before Any Action
University-scale Institutional AI
Note: This project demonstrates that responsible AI deployment doesn't mean avoiding AI — it means designing systems where AI handles what it's good at (pattern recognition, classification, structuring) while humans retain authority over decisions that affect people's lives.