IT Service Management
ITSM Ticket Intelligence
Real-time classification and routing for incoming IT support tickets, backed by a knowledge base and a chatbot that resolves the simple ones before a human sees them.
The problem
A global IT services company needed to cut the time it took to resolve IT service tickets. Every incoming ticket was triaged by hand, which meant the delay before work started was often longer than the work itself — and the same handful of simple issues kept arriving as fresh tickets.
What I built
I started by analysing years of historical ITSM data to understand what was actually arriving and who resolved it well.
From that, several pieces:
- Custom NLP models for information extraction, pulling the structured signal out of free-text ticket descriptions.
- A machine learning classifier that routes incoming tickets in real time by technical domain and criticality level.
- A data lakehouse consolidating the historical ticket data into a queryable knowledge base.
- A chatbot that resolves simple issues directly against that knowledge base, so they never become tickets.
- A near-real-time dashboard for monitoring how the models were performing in production.
The outcome
- Sharply reduced response and resolution times for IT support services.
- Fewer tickets overall, as the chatbot absorbed the common and simple issues.
- Surfaced which people were best suited to which categories of issue, improving how the IT department allocated its staff.
- Made the most common problems visible at a granular level — turning ticket volume into a signal about what to go and fix upstream.