Case Studies - Invoice Processing

Invoices are inherently complex and variable. Reconciliation of invoices is the process of extracting relevant information and storing them in a structured format to be easily tracked and managed.

Currently, organizations are heavily dependent on OCR techniques to automatically extract text out of the digital image of an invoice. While it is a step ahead of manual invoice processing, there is a fundamental reason why it is not considered a complete solution to the problem at hand.

Invoices from different sources tends to differ. There is no guarantee that critical information such as the dollar amount or invoice number will appear exactly at the same place across different documents. In other words, a typical OCR based solution does not always know where to 'look' to find the necessary information. Some organizations tackle this problem by creating a large number of static templates to address each invoice layout. This works as long as the number of layouts is small. It is also not a sustainable solution because an organization cannot keep on building new templates for every new layout.

The goal here is to build a solution that can streamline the above process by automating invoice extraction.