Sheet 01 / 03Illustrative example
AI triage for a growing support desk
Challenge
Support tickets were sorted by hand, so urgent issues waited in the same queue as simple questions and first replies were slow.
What we built
We built an LLM classifier that reads each ticket, tags product area and urgency, and routes it, with a person reviewing anything it is unsure about.
Results
- Faster first reply
- Most tickets routed automatically
- Human review on low confidence
- Client
- B2B SaaS company (under NDA)
- Industry
- SaaS
- Services
- AI DevelopmentCloud & DevOps
- Year
- 2025










