extract-data-invoice

Business Document Processing Extracting Fields from External Documents

In Automated Document Processing by Anandkumar S Nair

About the Author

Anandkumar S Nair

Advisory Consultant (Digital Products & Services). Anand S Nair comes with extensive hands-on experience in the software industry. He has proven track record in SAP & Non-SAP platforms. He has gained technology expertise with Oracle & other client-server technologies too. Anand supports & leads IVL, as the Advisory Consultant for Digital Products & Services.

Business Enterprises receive many kinds of documents related to supply chain, procurement and business administration from external agencies. These include:

  • Invoices
  • Purchase Orders
  • Bills of Entry
  • Import Licenses
  • Shipping Notes
  • HR Forms
  • …..
  • Traditionally, these documents used to be printed on paper and delivered through post or courier. Today, these come mostly as PDF files or as scanned documents attached to email. PDF (short for Portable Document Format) documents have now emerged as the most used solution for exchanging business data, particularly with trading partners.

    The main advantage of PDF documents is that these when printed out, retain the exact appearance and dimensions. Also text, images, tables etc can be embedded within these. PDF Documents can also be digitally signed.

    The trade-off in using PDF documents is that it is not easy to automate the extraction of text content and relevant field values from within these. This is particularly so when the text within a PDF document is embedded as an image, rather than as text.

    Text within a PDF document can be “copy/pasted”, or extracted using software tools – but this only when this is embedded as “text”, and not scanned as “image”.

When the document is either in PDF format with text content embedded as image, or if the document was scanned as an image in a format like JPG, then the text content within these can be extracted only using special-purpose software capable of recognizing text characters (alphabets, numbers etc) from within an image. Such tools are known as “OCR” (Optical Character Recognition) Software.

What is more, extracting the text from within documents is only part of the problem. How do we automate the extraction of the relevant fields? Say, in case of an Invoice, we would need to extract the Vendor’s name, GST Number, Invoice Number, Invoice Date, Reference Purchase Order, Names, Quantities, Rates line items etc. For this we need to use the software techniques like “natural language processing” and “machine learning”.

Innoval’s solution for document processing not only extracts relevant fields from PDF (and from scanned images), but is also capable of posting these directly into your ERP System like SAP – after due verification and authentication as dictated by the work-flow as adopted by the organization.

Innoval’s iDeck is the perfect alternative to hiring people for manual data entry from external documents to the business software of accounting systems used by businesses.