An invoice can arrive in a spreadsheet with every visible number intact and still produce the wrong total. Decide which values belong to each line before automating the import.
Prepare the fields before connecting the export
Parseur documents a PDF-to-Google-Sheets workflow for extracting fields and repeating table rows. Its described flow also handles scanned documents through OCR.
Create two field groups for an invoice. The first describes the document, such as invoice number and total. The second describes each line item, such as quantity and line amount.
Choose a stable invoice identifier before testing recurring imports. A filename alone may change when someone downloads or forwards the same document.
Set up the documented route
- Create a Parseur mailbox for the document workflow.
- Upload a synthetic invoice or forward it to that mailbox.
- Configure the header fields and repeating line-item extraction.
- Inspect the extracted data before enabling recurring delivery.
- Open the export options and select the documented Sheets route.
- Check the resulting rows in the destination spreadsheet.
The integration page describes a ready-made formula for the Sheets export route. It also identifies Zapier and Make as options when a workflow needs additional steps or mapping.
Confirm the current document allowance for the account you plan to use. A free test allowance does not establish the cost of processing a month’s invoices.
Use a small invoice with an answer you can check
The following fictional invoice, DEMO-001, is a validation fixture. It describes no real customer or purchase.
| Line | Quantity | Unit price (USD) | Line total (USD) |
|---|---|---|---|
| Notebooks | 2 | 12.00 | 24.00 |
| Pens | 3 | 2.00 | 6.00 |
The expected output has two line items. Their amounts add to $30; adding the $3 tax produces $33. Check those relationships after extraction and again after the data reaches Sheets.
An OCR error can change a digit without making the row look obviously broken. A quantity-times-unit-price check catches a different failure from a check of the final invoice total.
Keep invoice totals out of line-item sums
If the $33 invoice total is repeated on both exported line rows, summing that column produces $66. The extraction may have copied the document total correctly while the spreadsheet calculation uses it incorrectly.
Use a separate invoice summary area, or define which row supplies the document-level total. Keep line totals and invoice totals in clearly named columns.
To prevent duplicates, record the invoice identifier and a stable line reference. Reimporting DEMO-001 should trigger the workflow’s chosen duplicate policy instead of silently creating another copy.
Test the exceptions before relying on recurring imports
- Try a scanned copy and compare it with the native PDF.
- Move a line item onto a second page.
- Change a decimal separator and inspect the interpreted amount.
- Submit the same invoice again.
- Check a missing field before accepting the imported rows.
These are proposed validation cases, not results from a Parseur account. Keep an exception visible until someone resolves it; an empty cell should not become an invented value.
The provider’s extraction claims do not establish an error rate for your documents. Acceptance should depend on the saved rows and arithmetic checks for the actual workflow.
Our Parseur directory profile covers the product generally. The Zapier review discusses the broader automation workflow.
Sources
Which invoice-level value would be unsafe to sum once per line item?
Documentation checked October 25, 2026. Prepared with AI-assisted research and drafting. Worked examples are illustrative; MustHave.ai has not run an authenticated product test for this guide.
