From Paper Cash-Deposit Forms to Scan-and-Done

Every day, retail shops deliver their takings, cash, coins and cheques, to the bank. Each shop used its own paper form listing the number of notes, coins and cheques and their values. The formats were inconsistent and the entries were handwritten.

For the bank, processing those forms was slow and error-prone. A clerk had to read someone else's handwriting, re-key the figures, count the money, and reconcile the two by hand. Every step was a chance for a mismatch that then had to be chased down. Worse, all of this happened at the busiest time of day, so the slow process created queues at the counter exactly when there was least slack to absorb them.

The bank needed the deposit to arrive in a structured, predictable form, and the counter step to verify itself rather than depend on manual reconciliation.

We built a Laravel and React web application that lets retailers enter their deposit in a structured, validated way, rather than on a free-form paper sheet. Each submitted report carries a QR code that holds all of the deposit's data.

At the bank, the process becomes scan-and-done. The clerk scans the QR code and feeds the cash into the counting machine. The counted amount is then cross-checked automatically against the figures the retailer declared, and the deposit is complete. There is no re-keying, no interpreting handwriting, and no mismatched totals to chase, because the declared data and the counted data are reconciled by the system, not by a person under time pressure.

Processing a deposit at the counter became four times faster. The structured digital form and the QR cross-check replaced inconsistent paper forms, so the numbers are right the first time and the queue keeps moving, even at the end-of-day peak.

The deeper change is where the work happens. Instead of a clerk reconciling two sources of truth by hand, the system reconciles the declared figures against the counted figures automatically. People stop transcribing and start verifying, which is both faster and far less error-prone.

The win here did not come from a clever algorithm. It came from moving the structure upstream.

The slow, error-prone part of the old process was a person at the counter trying to reconcile a free-form paper form with a pile of cash. By making the retailer enter structured, validated data at the source, and carrying it in a QR code, the hard reconciliation step downstream became an automatic check. The lesson we keep relearning is that the cheapest place to enforce data quality is at the point of entry. Fix it there, and most of the downstream pain simply disappears.