Skip to main content

Transaction matching

Agent Reconciliations uses AI-assisted matching to compare transaction activity from two sources and prepare a reconciliation for your review. It can match data from Excel workbooks, live transaction detail from your connected ERP, or one of each.

note

Transaction matching is part of the agent-task capability and may not be enabled for your account. Contact your account manager if you'd like access.

What it can match​

Set up a saved reconciliation for the two sides you need to compare. For example:

  • A bank statement against a GL cash account.
  • A Stripe payout report against the corresponding deposits.
  • A subledger against its GL control account.

For each source, CloseCore maps the fields that identify a transaction, such as amount, date, and reference. You can also filter the source data and allow matching transaction groups when one side of a reconciliation corresponds to several transactions on the other side.

How matching works​

The reconciliation configuration defines the criteria a match must satisfy and can score likely matches using the fields you choose. You can set a confidence threshold so only sufficiently strong matches are included.

When an agent runs the reconciliation, CloseCore writes the results into a workbook for review:

  • A Summary sheet with matched and unmatched counts.
  • A Matches sheet with the transaction groups that matched.
  • Separate Unmatched sheets for each side, so you can investigate the remaining items.

When the source data and results are in the same workbook, result rows link back to the original transaction rows.

Review and sign-off​

Transaction matching prepares the reconciliation; it does not replace review. Investigate any unmatched activity, then use CloseCore's normal reconciliation and sign-off workflow to complete the work.