Optical Character Recognition

Optical Character Recognition

  • Optical Character Recognition and Intelligent Character Recognition are used to replace the manual input task in document processing tasks. These solutions can process both image files and digital documents. They usually extract the text information into a plain text file or into a formatted electronic document that can be further used by other tools.

    Intelligent OCR solutions use machine learning algorithms or templates to identify documents and read out specific information like addresses, line items on invoices, ID’s, etc...

    OCR is typically a component in an automated workflow where other tools can complement the capabilities like workflow systems that can manage the approvals or RPA that can act as an interface between systems.

  • Typical use cases

    • Accounts payables process: Read invoices, extract necessary information to input to the ERP systems (e.g. SAP, E1, Navision). To read out specific information like line items, tax, sum, etc. a more complex, smart OCR solution is required where the implementation is longer and more expensive.
    • T&E audit process: Use OCR to scan uploaded invoice copies to find the claimed amount, dates, type of invoice and anything else that is needed to be audited. A basic OCR is sufficient for this purpose.
    • Order capture: Incoming purchase orders can be processed by scanning the image, identifying the ordered items that can be populated to an ERP order using RPA.
    • Text Fairy or Google Lens which are standard mobile phone applications and can scan either transform the text on images to characterized output or can read in texts live through the camera.
    • Google Translate on mobile apps can not just read text, live through the camera but can instantly translate the text which the user is focused on.
  • Benefits

    • OCR/ICR can extend the traditional automation project scope where data input arrives in an „image” format (can be also a scanned pdf) and instant data extraction of the information is not possible.
    • As OCR is a standard built in feature of UiPath, implementation doesn’t require any costs
    • UiPath has many language packages built in, so language dependency is minimized to very special cases only.