Import the eBay file
The user uploads the weekly remittance file, 300 to 500 payment references at a time.
Accounts payable automation case study
eBay Auto AP Reconciliation.
A custom ERP module that takes a large weekly eBay remittance file, matches it to the penny, flags what doesn’t match, and posts hundreds of balanced entries with one click.
Custom ERP module designed and built by Built-N-AI founder Casey Jackson.
Results at a glance
~15 minPer remittance fileDown from half a day or more
+55%Accuracy increaseVersus the manual process
300–500Payment references per fileProcessed in a single weekly import
To the pennyEvery posting balancesMissing credits and invoices auto-created
1 clickUpdate buttonHundreds of postings in one action
Correct GLAccounts on every postingConnected to the right general ledger accounts
The challenge
Hundreds of payments, every week, none of them tidy.
Every week eBay sends a remittance file covering 300 to 500 payment references. Each one has to be matched to the right invoice, to the penny, and posted to the right general ledger accounts. Anything that doesn’t match — a short payment, a missing credit, an invoice that was never created — has to be researched and fixed by hand before the books will balance.
Done manually, a single file could take half a day or more, and every manual match was another chance for error.
“The system does the matching. The user makes the decisions.”
The solution
Upload the file. Review the exceptions. Click Update.
We built a custom module inside the ERP that imports the eBay remittance file and does the reconciliation work. Exact matches are confirmed automatically. Mismatches are surfaced with a confidence score for the most likely correct match, and the user approves it or changes it on the fly.
Missing credits and invoices are created automatically so every posting balances to the penny. One click on Update posts the whole file.
What the user sees
- Exact matches — one to one, to the penny and invoice number
- Mismatches — each with a confidence score
- Approve or change any suggested match on the fly
- Missing credits and invoices — created automatically
- Balanced postings to the correct GL accounts
Import → Match → Review Mismatches → Auto-Create Missing → Update.
A human-in-the-loop design: the module handles the volume, and the user only spends time on the exceptions that need judgment.
How it works
Five steps, one file.
Exact matches reported
Every payment that matches an invoice one to one, to the penny and invoice number, is reported back to the user.
Mismatches scored
Each mismatch is reported with a confidence score for the match the module believes is correct. The user approves it or changes it on the fly.
Missing items auto-created
Missing credits and invoices are created automatically so the posting balances to the penny.
One click to post
The user clicks Update. Hundreds of postings, new invoices and credits are posted and balanced at the end.
Built through iteration
The edge cases are the real work.
Reconciling a clean payment is easy. The difficulty is everything else. This module took many iterations to account for the edge cases in real remittance data and to decide how each one should be handled, so that the output still balances to the penny.
Every posting is connected to the correct GL accounts, so the accounting is right the moment the user clicks Update, not after a cleanup pass.
“Automation handles the volume. People handle the judgment calls.”
Before & after
The reconciliation model changed.
Before
Half a day or more to reconcile a single file
After
About 15 minutes per file
Before
Each payment matched to its invoice by hand
After
Exact matches identified automatically, to the penny and invoice number
Before
Mismatches researched one at a time
After
Each mismatch arrives with a confidence-scored suggestion to approve or change
Before
Missing credits and invoices created manually
After
Created automatically so the posting balances to the penny
Before
Postings entered one by one
After
Hundreds of postings made with one Update click, to the correct GL accounts
Business impact
More accurate books in a fraction of the time.
Per file
What could take half a day or more is done in about 15 minutes.
Accuracy
Fewer matching and posting errors than the manual process.
Balanced postings
Missing credits and invoices are created so every posting balances.
GL coding
Every posting lands in the right general ledger accounts.
Review time
The user spends time only on the mismatches that need a decision.
Key takeaway
Hundreds of payments, reconciled and posted in minutes.
The module does the matching, scoring, creating and balancing that a person used to do by hand, and leaves the judgment calls to the user.
The system does the matching. The user makes the decisions.