BLOG

Intelligent automation in your administration

Intelligent Document Processing (IDP).

What is Intelligent Document Processing (IDP) ?

The term IDP is increasingly used by software vendors to describe the steps required to extract data from documents using artificial intelligence (AI). It is the process of translating the content of a document (physical or electronic) into something meaningful that a machine (computer) can process. It is the process of translating the content of a document (physical or electronic) into something meaningful that a machine (computer) can process. The process converts the incoming data into a “structured” output so that automated software such as RPA (robotic process automation) can take over and perform automatic input into various systems.

How does intelligent document processing work?

Unlike conventional OCR, IDP can recognise much more than just characters. It is – of course – also capable of interpreting those characters and uses different aspects of AI, such as natural language processing (NLP) and machine learning (ML), to help classify data and group the same categories of data.

The best solutions are also already a far cry from the accusation sometimes made that they are model-based and therefore labour-intensive. This is no longer the case! Documents are examined against templates created according to the type of document.

The ability to recognise nested tables or rules and convert them into structured data is also an important differentiating factor. Find out more about this, as not all suppliers are equally strong in this area!

Won’t accelerated digitalisation make IDP obsolete?

The flow of paper documents will not disappear any time soon, regardless of the digitisation initiatives. This is mainly due to a difference in digitisation speed and cost between companies. Not all companies will go paperless at the same speed. This means that a single uniform process is an illusion. Another important aspect is that it ignores the possibility of involving a human being in the process. A human being can read/interpret pdf’s, but xml or json are much more difficult to read.

Solutions such as the UBL standard even include a field to convert the PDF to BIN64 format, which makes the PDF visible again in an application’s user interface. In summary, at the moment I do not (yet) believe in the full digitalisation of our economy.

Which companies will benefit from an IDP solution?

All companies with a large paper flow still need an IDP solution.

Choose the documents that still flow through your company from the list below:

  • commercial documents
    • bills of lading
    • airway bills
    • letters of credit
  • Documents for creating new customers or suppliers
  • Financial documents
    • invoices
    • offers
    • bank extracts
  • Certificates
    • diplomas
    • certificates of origin
  • documents relating to stock management
    • packing lists
    • delivery notes

With IDP, people no longer have to manually process large quantities of documents. By converting incoming data from PDFs (whether scanned or native PDF documents) into Excel/CSV, XML, JSON files, workflows become much easier to automate later, e.g. through the use of software robots (RPA – Robotic Process Automation).

Table of contents