01 / THE BLUEPRINT
The problem to investigate
A shop taking several payment methods may struggle to connect a sale, a provider reference and a settlement. Interview merchants about actual mismatches rather than assuming fraud. Establish whether the problem is record-keeping, delayed settlement or an existing tool used incorrectly.
02 / THE BLUEPRINT
A possible solution
Build a review queue that matches imported sales and authorised provider records by reference, amount and date. Show unmatched items for a person to investigate. Keep the product as a bookkeeping aid; it must not receive customer money or decide a payment succeeded from a screenshot.
03 / THE BLUEPRINT
Start without an app
Try a daily spreadsheet with sale identifiers, verified payment references and a sign-off routine. Separate pending transactions from confirmed receipts. If a merchant’s existing point-of-sale export already answers the question, training may be enough.
04 / THE BLUEPRINT
Technology that could help
Begin with validated CSV imports and a deterministic matching routine. Later, authorised provider APIs or signed webhooks may help where supported. Encrypt stored data, minimise identifiers and retain a visible audit trail for manual adjustments.
05 / THE BLUEPRINT
Your first experiment
- Ask one willing merchant to describe a recent reconciliation problem without sharing customer secrets.
- Create a synthetic sales and payment dataset containing duplicates, reversals and delayed entries.
- Have a bookkeeper reconcile it manually, then compare the prototype’s suggestions.
- Run a permissioned, read-only pilot only after security and data-handling review.
06 / THE BLUEPRINT
People, resources and cost drivers
You need merchant workflow knowledge, bookkeeping input and integration skills. Costs include provider integration maintenance, secure hosting, support and export handling. Estimate these from actual transaction volumes; a small pilot does not establish a viable subscription price.
07 / THE BLUEPRINT
What to validate — and when to stop
Validate duplicate handling, refunds, timezone differences and settlement timing. Stop automated matching if references are unreliable or it hides exceptions. A false match can be worse than an unmatched line, so keep a human confirmation step and an easy correction path.
08 / HYPOTHESIS TO TEST
Potential impact and how to measure it
The proposed benefit is less time closing the books with fewer unresolved differences. Compare reconciliation minutes, false matches and remaining exceptions. Do not describe flagged differences as recovered revenue or evidence of theft without a separate investigation.
09 / THE BLUEPRINT
Risks and safeguards
Payment records can expose customer information and commercial activity. Avoid collecting wallet PINs, login credentials or unnecessary account details. Do not scrape private accounts or assume a provider permits every integration. Check local requirements before expanding beyond a read-only record tool.
10 / THE BLUEPRINT
Who might pay or maintain it
Merchants or bookkeeping services might pay for reliably reduced administrative work. Test willingness to pay after a useful pilot, not before proving the workflow. Preserve exports so the shop can leave without losing its own records.