Project 02
AI IT Help-Desk Incident Triage Agent
An AI-assisted application that classifies support tickets, creates structured incident summaries, and recommends practical routing and next steps.
01 / Overview
What it does
This AI-assisted help-desk application explores how structured AI output can support consistent incident triage. It analyzes a submitted ticket and returns an organized assessment for a support workflow.
02 / Context
Problem
Incoming support tickets often vary in detail and urgency. Early classification can take time, and inconsistent summaries make routing and escalation harder.
Solution
The application guides an AI model to return structured fields for category, severity, priority, summary, support-team recommendation, and suggested next actions.
03 / Capabilities
Features
- Accepts IT support ticket text
- Classifies category, severity, and priority
- Generates concise incident summaries
- Recommends an appropriate support team
- Suggests troubleshooting or escalation steps
- Recognizes common security-related incidents
04 / System design
Architecture
05 / Reflection
What I learned
- Designing predictable structured AI responses
- Mapping IT and security concepts into a triage workflow
- Testing classifications with varied ticket examples
- Presenting AI output without overstating certainty
Engineering considerations
- Inputs are treated as untrusted text.
- AI recommendations are guidance for review, not automatic incident resolution.
- The project does not claim enterprise production use.
06 / Next steps
Future improvements
- Add a larger test set for category consistency
- Introduce confidence-aware review states
- Add optional ticket-system integration
- Expand accessibility and error-state testing