Fintech · 2025 · Client project · Live
ChamaSmart
No more manual reconciliation for savings groups.
Visit the live site - 25
- Financial rules applied automatically
- Integrated
- M-Pesa payments
- 5
- Health score measures
The problem
Spreadsheets and chat messages. Committees lose hours reconciling and chasing loans.
What we built
M-Pesa contributions straight in. Rules score each member and group. An AI assistant answers from the group's own history.
What it does
- 25 rules that produce a personal financial health assessment
- Five-measure group health score: contributions, repayment, engagement, growth and risk
- Contributions collected through M-Pesa
- AI assistant that uses contribution and loan history
- Automatic reminders, overdue alerts and monthly reports
How it was built (technical)
- Next.js front end with Firebase sign-in
- Node.js services on PostgreSQL, with Redis for fast reads
- Background workers for reminders, overdue loan alerts and monthly reports
- A rules engine that scores financial health on five measures
- Live updates to the group health score
- Next.js 14
- Node.js
- PostgreSQL
- Redis
- BullMQ
- Socket.IO
- Groq AI
- M-Pesa Daraja v2
- Firebase Auth
Client: KCA University. Results shown are those reported for this project.