Multi-Agent IT Helpdesk
A production multi-agent system on AWS Bedrock serving 1,500 employees. Deflects ~40% of the L1 queue and answers in seconds instead of hours.
Role
Product Owner & AI Lead
Timeline
2025 · 8 months to production
The challenge
IT was drowning in repetitive L1 tickets, password resets, VPN issues, software access, while real problems waited behind a week-long average resolution time.
The decision
Chose Human-in-the-loop over Full auto-close
A wrong auto-close costs more trust than the minutes it saves.
How I approached it
- 01Mined ticket history to find the repeatable intents worth automating; wrote the business case and PRD with deflection and resolution time as north-star metrics.
- 02Designed a multi-agent architecture: a parent agent routing via Azure AD to an employee self-service bot and an IT-support bot, each with its own knowledge and permissions.
- 03Built RAG on AWS Bedrock grounded on IT knowledge and Matrix42 ticket data, with PII redaction in the ingestion pipeline.
- 04Ran governance inside the build, GDPR, EU AI Act classification, works council, and shipped with escalation paths plus feedback logging.
Impact
- ✦~40% of L1 tickets deflected in year one.
- ✦Average resolution time from 1 week to 2 days.
- ✦~€100K/yr saved in working hours across 1,500 employees.
- ✦Tickets that reach IT arrive pre-filled with conversation context.
Stack
- AWS Bedrock
- Claude
- S3 / Lambda
- Strands Agents
- Azure AD
- Python