Skip to main content
Many brands manage retailer deductions (chargebacks) through a manual process: download an invoice PDF from a retailer portal, convert it to Excel, add lookup columns to map codes to internal categories, pivot to summarize, and reconcile against a contract. This is a strong fit for Parabola because it combines PDF parsing, a lookup join, and aggregation — all steps that repeat every time a new invoice arrives.

​
The manual process this replaces

A typical pre-Parabola workflow looks like this:
  1. Download an invoice PDF (or backup file) from a retailer’s vendor portal.
  2. Convert the PDF to Excel.
  3. Add a derived column — for example, extracting the last 5 digits of a UPC with a formula like =RIGHT(I2,5).
  4. Add a lookup column (VLOOKUP) mapping that derived value against a reference table to bring in an internal category or product group.
  5. Build a pivot table to summarize by category.
  6. Manually validate the summary against the contract terms and reconcile.

​
The Parabola equivalent

1

Pull and parse the invoice

Use Extract from email to ingest the invoice as an email attachment, or Extract from PDF for a one-off upload. Configure header fields (invoice number, date) as keys and the line-item table as columns. For scanned or inconsistently-formatted invoices, switch to OCR-first parsing in the PDF advanced settings.
2

Recreate any Excel formula columns

Replace formulas like =RIGHT(I2,5) with reproducable and easy-to-use AI steps.
3

Recreate the VLOOKUP as a join

Pull in your lookup/reference table (a UPC-to-product-group mapping, for example) and use Combine tables to join it to the invoice line items on the derived key column.
4

Create clean and stunning dashboards to review

Use Prowork ro create visualizations to view and share your data.
5

Format and export

Match your target export layout, then send the result to email, a spreadsheet, or your destination system with the appropriate step.

​
Before you build

To get the mapping right on the first attempt, gather:
  • The exact columns and order needed in the final export.
  • Where the export should land — email attachment, cloud drive, API, etc.
  • A sample of your lookup/reference table headers (for example, the UPC column name and the category column name it should resolve to), so the join keys line up correctly.
Last modified on July 14, 2026