Skip to main content
When a new bank transaction is imported, Upflow automatically analyzes it and tries to identify the most likely customer and invoice match. If a suggestion is found, it is surfaced directly in the Bank transactions list and pre-filled in the matching interface — so you can approve it with a single click instead of searching manually.
When the match is clear, the Cash App agent can skip the review step and apply the transaction for you automatically.

Identify transactions with suggestions

In the Bank transactions list, the Suggestions column shows a badge with the number of suggested invoice matches for each transaction. The Bank transactions list showing the Suggestions column with a +1 badge on each row, indicating one match suggestion per transaction. A badge of +1 means Upflow found one suggested invoice match. Transactions without a badge have no suggestion and must be matched manually.

Review and approve a suggestion

  1. In the Bank transactions list, click the Apply button on a transaction that has a suggestion badge. The matching interface opens with the Suggested tab active. The suggested invoice is listed, along with the matched customer and the invoice amount. The matching interface modal open on the Suggested tab, showing one pre-populated invoice row with customer name, due date, status, and outstanding amount.
  2. Click the → arrow button on the suggested invoice row to fill in the outstanding amount. Upflow auto-fills the customer dropdown and the amount to apply, and moves the invoice to the Selected tab. The footer updates to show the total selected and remaining amount. The matching interface after clicking the arrow button: the invoice is checked, the customer is auto-filled, the amount field shows the full outstanding amount, and the Apply fully button is enabled in the footer.
  3. Click Apply fully to apply the transaction.
    👉 If the suggested amount does not match the full transaction amount, the button will read Apply partially instead.

Override or ignore a suggestion

If the suggestion is incorrect, you can switch to the Invoices tab and select the right customer and invoices manually. The suggestion is not applied unless you click Apply. Switching tabs does not delete the suggestion — if you come back to the Suggested tab later, the original proposal is still there.

How suggestions are generated

Upflow runs a matching algorithm each time a new bank transaction is imported. Suggestions are built using the following information:
  • Invoice references in the transaction description. Upflow recognizes an invoice number even when the reference doesn’t match it exactly.
  • The payer name, matched against your customer names. Upflow still finds the customer when the name is truncated or spaced differently from the one in Upflow.
  • Your previous cash application history for the same payer.
  • The transaction amount, compared against the customer’s outstanding invoices.
The higher the quality of the match, the more confident the suggestion. If the customer sent a remittance advice for the payment, Upflow uses it first and suggests the invoices it lists.
When Upflow finds the customer from an approximate name match, the invoice stays a suggestion for you to review, even if the amount matches exactly. The Cash App agent doesn’t apply it automatically.

Partial payments

When the transaction references a specific invoice but pays less than its outstanding amount, Upflow still suggests that invoice. Apply it with Apply partially.

One payment for several invoices

When a transaction has no invoice reference and its amount doesn’t match any single invoice, Upflow looks for a combination of the customer’s open invoices that adds up to the transaction amount. If several combinations work, Upflow suggests the one with the oldest invoices.