What it does
A reconciliation automation for a CA firm. Each month, every client's purchase register is matched against the GSTR-2B from the GST portal. Results fall into four groups: matched, amount mismatch, in books but not in 2B (input credit at risk) and in 2B but not in books (credit not yet claimed). Near-matches with typos in invoice numbers are suggested for review. The article clerk reviews the suggestions, supplier follow-ups are drafted, and the senior signs off before filing.
The problem
For a client with a few hundred invoices, matching by hand took two to three hours a month, and with dozens of clients the firm spent days on it every month. Credit at risk was found late, after the supplier had stopped responding.
Approach
Matching looks at GSTIN, invoice number, date and amounts together, so "INV-091" and "INV91" are suggested as the same invoice. Every suggested match is shown to a person, never accepted silently. Supplier follow-ups are drafted with the exact invoices missing, and the senior approves the final reconciliation before returns are filed.
- Input
- AI step
- Rules or code
- Person decides
- Result
Screens
Swipe sideways to see the whole screen.
Client work is shown under confidentiality agreements: no client names, and screens are recreated with sample data.
Stack
- GST portal data
- Python matching
- Gemini API
- Email follow-ups
Impact
Typical for systems like this.
