FinSave AI: what it takes to build financial coaching for MSMEs
FinSave AI is live at finsave.aitoolshq.space, built around bank and mobile-money coaching for small businesses in Rwanda.
The problem it's aimed at
Most MSMEs in Rwanda don't have a single bank account they run their business through — money moves across mobile money, informal record-keeping, and sometimes a bank account, without ever being aggregated in one place. FinSave AI is built around that reality rather than assuming a single clean bank feed.
The ingestion approach
The planned approach for pulling in financial data combines Plaid Sandbox for bank accounts with a mobile-money statement upload path — accepting CSV, PDF, or even a photo of a printed statement, run through Tesseract OCR to extract the transaction data. [PLACEHOLDER — confirm current implementation status: this is the planned approach from project docs; state plainly here whether ingestion is fully live yet or still in progress, since the fetch only confirmed the site resolves, not what's functionally complete.]
Lender access without exposing real account numbers
One deliberate design choice: lenders reviewing a business's financial standing don't see real account numbers. Instead, FinSave generates a separate Business Code used specifically for lender-facing access — keeping the underlying account details out of a channel that doesn't need them.