← All work
Kira
AI support assistant that resolves 62% of tickets by grounding answers in the client’s own docs.
The challenge
A fast-growing SaaS team was drowning in repetitive support tickets, and a generic chatbot had made customers more frustrated, not less.
What we built
- 1
Retrieval over the product docs, changelog and past tickets so answers cite real sources.
- 2
An evaluation suite of 600 real questions run on every change before release.
- 3
Smooth hand-off to a human agent with full context when confidence is low.
- 62%
- tickets auto-resolved
- −48%
- first-response time
- 91%
- answer accuracy
Stack
- Python
- LLM APIs
- pgvector
- FastAPI
- Next.js
Sample case study. Swap in real screenshots, numbers and a client quote before launch.