Splitting Invoices in ABBYY Vantage

There are many cases when more than one invoice is contained in the same file. Sometimes vendors will send the invoices in this manner and other times the company may scan several images together into one output file. In this scenario, it is important to split the invoices into separate objects for processing, as they are unique content items and need individual processing/approvals.

ABBYY Vantage can automate this using a Splitter skill. This skill is included in the Process skill as part of an Assemble step. When included with the M-Files integration with Vantage, the process will work as follows:

1. Multi-invoice document is added to M-Files
2. The document is sent to ABBYY Vantage for processing
3. The document enters the "Assemble" step in the Process skill
4. Using the Splitter skill, a single document is split into several invoices
5. Each invoice goes through the rest of the Process skill
   1. data extraction and, if required, user review
6. When the invoices are complete in ABBYY Vantage and synced back into M-Files, each individual invoice is added as its own object in M-Files

Splitter Skill

A Splitter skill is a type of skill created in ABBYY Vantage. Using a bit of script as part of this skill, you can easily split the documents into multiple single-invoice files. While there are many ways to achieve this, for the purposes of this article, the logic is as follows:

• Create a Splitter skill using ABBYY Vantage Advanced Designer
• Copy the Invoice Document Splitter Skill provided by ABBYY
• Open the skill for editing

Click on the Splitter Script step in the skill and then Script Editor (lower right portion of the screen). From here update the script to work as needed – the logic deployed for this example is as follows

• Loop through all the pages in the document
• Set the invoice number, if found on the page
• If the invoice number on the current page is different from the previous invoice number
  • This is the first page of a new invoice document
  • Push the current page to a new document
• If the invoice number is null or the same as the previous page
  • Push the current page as a subsequent page to the current document
• Return the array of documents to the calling (process) skill

Now we have several invoice documents where they used to be a single multi-invoice document. From there the Process skill takes over and treats each document as a separate entity.

Process Skill

The invoice process skill has the following steps to accomplish this task:

• Input – file is received by ABBYY Vantage
• OCR – document is converted from an image to full text searchable
• Assemble – the document is split into one too many invoice documents
• Extract – data is extracted from the document into the form fields
• Review – optional user review
• Output – document and captured data is sent to M-Files

The Assemble step is where the process references the splitter skill – done by clicking on the Assemble step and selecting ‘By Document Splitter Skill’, and then selecting the appropriate splitter skill.

Reach out to us here at TEAM IM for further information on automating document processing in M-Files or for how to add invoice splitting to your ABBYY Vantage processes!