GramUdyam Saathi
Voice-first loan advisor for first-time rural entrepreneurs under MoSJE schemes. Twelve questions in Hindi or English, then a right-sized loan — deterministic engines for scheme routing, amortisation, DSCR, catchment feasibility and a 1,000-run Monte Carlo, plus VLE and officer dashboards.
Borrowing the right amount
First-time rural entrepreneurs under MoSJE schemes often borrow what a formula says, not what their business can repay. GramUdyam Saathi asks twelve questions, in Hindi or English, and recommends a loan they can actually service.
Real maths, not a chatbot
The engines are deterministic code: scheme routing with edge cases, quarterly amortisation after moratorium, DSCR and break-even, catchment feasibility, and a seeded 1,000-run Monte Carlo for default risk.
- Recommends the largest unit with DSCR ≥ 1.5 and default risk ≤ 15%
- Golden test cases run in the browser
- A numeric guard rejects any figure not backed by the facts
Built for every desk in the chain
Beneficiaries get a voice-first flow, village operators get a fast keyboard form, and scheme officers get a case queue with evidence and a bank-format project report.
Results
Questions people ask
Is the data real?
The prototype runs on demo data for Barabanki district shaped after public datasets; the finance and feasibility logic is real.
