Business

From PDF Zone to a Custom AI-Assisted Invoice Renamer

/ John Wagenvoort / Leave a comment

For many years our supplier invoice renaming workflow relied on PDF Zone.

PDF Zone did a good job when dealing with a handful of suppliers that followed consistent invoice formats. We created supplier profiles and used those profiles to rename invoices before uploading them into EzzyBills.

As our supplier base grew, however, the limitations became increasingly apparent.

Every new supplier required a new profile. Invoice layouts varied significantly. Some suppliers embedded text in their PDFs, while others supplied image-based invoices that were difficult to process reliably. Maintaining individual supplier profiles became a never-ending task.

Earlier this year we decided to explore whether modern AI-assisted development tools could help us build something tailored specifically to the way Rosemount Joinery operates.

The result is the Invoice Renamer.

What the Application Does

The application scans an entire folder of supplier invoices and automatically identifies:

  • Supplier name
  • Invoice number
  • Job reference
  • Service provider invoices
  • Special supplier-specific numbering formats

It then proposes a standardised file name before the invoices are uploaded into EzzyBills.

Typical naming formats include:

Vendor – Invoice Number – Job Reference.pdf

or

Vendor – Invoice Number.pdf

for service providers where no job reference applies.

Instead of processing invoices one at a time, the application presents all invoices in a review table where proposed names can be checked before any files are renamed.

OCR Support

One challenge encountered during development was that many supplier PDFs contain little or no searchable text.

To address this, the application now uses multiple extraction methods:

  1. Embedded PDF text extraction
  2. PDFKit text extraction
  3. Apple Vision OCR

This allows scanned invoices to be processed without modifying the original PDF files.

Built Around Real Supplier Invoices

Unlike generic renaming utilities, the Invoice Renamer has been trained against real supplier invoices used every day within our business.

Support currently exists for suppliers including:

  • Polytec
  • Hafele
  • Lincoln Sentry
  • Laminex
  • Wilson & Bradley
  • Hays
  • Greg Steele
  • Verdecor
  • Bord Products
  • Surteco
  • Elegant Hardware

and many others.

Supplier-specific rules can be maintained separately from the application itself, allowing new suppliers to be added without rebuilding the software.

Manual Override Capability

One of the most recent additions has been the ability to manually edit invoices that belong to unusual or infrequently used suppliers.

Rather than creating permanent supplier rules for a supplier that may only appear once every year or two, the user can simply edit the proposed filename directly within the application.

This provides flexibility while keeping the supplier rules database manageable.

Native macOS Application

What began as a collection of Python scripts has evolved into a native macOS application.

The application now includes:

  • Folder settings
  • Preview table
  • Quick Look PDF preview
  • Editable proposed filenames
  • Rename workflow
  • Rename and Move workflow
  • CSV reporting
  • Supplier and reference management

The objective has always been simple:

Reduce manual processing while allowing users to maintain complete control over the final result.

Introducing Source Control

Another milestone achieved during development was the introduction of GitHub source control.

Historically, development moved between multiple Macs using manual folder copies and backups. While effective, this approach made it difficult to track changes over time.

The project now uses GitHub to:

  • Maintain version history
  • Synchronise development between multiple Macs
  • Provide rollback capability
  • Record development milestones

Traditional backups remain in place, but GitHub now provides a structured development workflow.

Looking Ahead

The project continues to evolve.

Future development will focus on:

  • Enhanced supplier training
  • Duplicate invoice detection
  • Improved reporting
  • Additional workflow automation
  • Broader deployment within the business

What started as a replacement for PDF Zone has become a purpose-built business application designed specifically for the way Rosemount Joinery processes supplier invoices.

Most importantly, it saves time, reduces manual effort, and allows our administration team to focus on more valuable work than renaming files.