Voice-first agentic AI for worker-verified job impact in Ethiopia.
# Mela Verify
> **Every worker has a voice.**
Mela Verify is a voice-first, AI-assisted worker verification platform designed for employment programme monitoring in Ethiopia.
Instead of relying only on employer-reported monitoring forms, Mela Verify gives workers a direct and accessible way to describe their own employment conditions in **English or Amharic**.
The system converts simple worker conversations into structured evidence that programme monitoring teams can review alongside employer-reported data.
---
## The Problem
Employment programmes often measure impact using information reported by employers.
But an important voice can be missing:
**the worker.**
Workers may have information about:
- how long they have actually worked
- how many hours they work
- what they are actually paid
- whether they are free to leave their job
- whether they experience discrimination
- whether they are free to organize with other workers
Traditional digital forms can also exclude workers who:
- have low digital literacy
- have limited internet access
- prefer Amharic
- do not own a smartphone
- find formal monitoring questions difficult to understand
Mela Verify was designed around one question:
> **What if the monitoring system could simply talk to the worker?**
---
## Our Solution
Mela Verify uses one interview engine through two worker-access channels:
### 📱 Smartphone Interview
Workers can complete a mobile-friendly interview using:
- voice
- text
- English
- Amharic
Mela asks **one simple question at a time**.
After the interview, the worker reviews their answers, can correct them, and confirms:
**“Yes, everything is correct.”**
Only after confirmation is the interview submitted to the monitoring system.
### ☎️ Feature-Phone Experience
Mela also demonstrates how the same interview could work through a basic phone / IVR experience.
For the hackathon MVP, the telecom layer is **simulated in the browser**.
The worker can:
1. dial the demo number …