A customer returns one item from a two-product promotion. The original order used loyalty points, a gift card, and a bank card. Delivery was free because the basket passed a spending threshold.
What is the correct refund?
The answer may depend on how the discount was allocated, whether loyalty points must be restored, which payment method receives the money, and whether removing the item makes the original order ineligible for free delivery. The warehouse can confirm that the product came back without resolving any of those questions.
This is why retail returns reconciliation often remains manual. Matching a returned SKU to an order is only the beginning. The retailer must also reconstruct the commercial logic of the original sale.
The NRF estimated that 15.8% of annual US retail sales would be returned in 2025, representing $849.9 billion in merchandise.
The selling price is rarely the complete financial record
Modern retail transactions contain product discounts, basket promotions, vouchers, loyalty rewards, shipping fees, tax, duties, marketplace commissions, and several payment methods.
A return changes some of these values while leaving others intact. Shopify’s documentation reflects this complexity. Return fees and exchange items affect whether a refund or additional payment is due, while split refunds may need to be assigned across separate payment methods.
A spreadsheet becomes the reconciliation layer when the commerce platform, payment provider, and ERP apply these adjustments at different levels. One system works at order level, another at payment level, and finance may require entries for each returned line.
Product condition changes the accounting outcome
The same jacket can return to available inventory, move to an outlet channel, require refurbishment, or become a write-off. Those outcomes affect inventory value, margin, tax treatment, and the timing of the accounting entry.
The refund may already be complete when the commercial value of the returned item is still unknown. Finance then holds an open difference until warehouse disposition or a manual accrual closes it. The broader retail returns process automation must therefore connect the financial outcome with the physical disposition of the product.
Manual review often protects policy decisions
Retailers also need judgment when a return involves an expired policy, missing receipt, suspected abuse, partial bundle, marketplace seller, or cross-border purchase. The reviewer is interpreting commercial policy rather than correcting a simple data error.
Robotic process automation in finance can match references, compare amounts, and prepare exception files. AI enterprise automation becomes useful when it applies promotion, payment, tax, and disposition rules consistently across channels while retaining approval controls.
The strongest automation candidate is a repeated reconciliation reason with a documented decision rule. Start with split refunds, promotion allocation, loyalty reversals, or inventory write-downs. Each resolved rule removes a specific class of manual investigation without hiding the financial decision behind the workflow.
