A New Way to Look at OCR

3 Key Differences Between Traditional OCR and ktIQ

Optical Character Recognition (OCR) in simple terms is the ability of software to convert images into readable text. The images are typically received from a scanner, an email, or other file upload or import.

The purpose of OCR is to make the files easy to store/archive, retrieve data from and search fields or specific words in those files. Ultimately, the documents are saved in a document management system in a common file format, i.e. PDF, Word, Excel, or even in their native image format, making it easy for business users to find files when they need them.

In addition, when used in combination with a workflow process, it is common for the OCR system to receive the file and extract some data from the file and populate key fields about that document in order for business users to determine how that file should be managed.

For example, let’s look at an invoice workflow process that begins with OCR. The goal of the software and process automation is to:

  • Receive an invoice in a digital format, even if it was sent on paper and scanned into the system
  • Extract some key data from the invoice and populate the fields in an invoice format
  • Based on the invoice data, send the invoice to a member of the finance team or a budget owner/manager for review, edit or approval
  • Once the invoice is approved, send the invoice, along with any comments and feedback from the reviewers, back to accounting for final processing
  • Often the final processing involves ERP system integration so that the entire process is automated

The ideal solution in this process would accurately extract the data and enable the user or system to send the invoice to the appropriate reviewer with very little, if any, data entry or correction. This saves a lot of manual effort for data entry, reduces the risks of error, process delays, and streamlines the process with system-generated email notifications at each step.

Now, when evaluating accounts payable invoice automation solutions, particularly those that include OCR on the front-end of the process, it’s important to understand the full scope of what’s required to implement the software and configure it correctly to work with your vendor invoices.

All invoices are NOT created equal. The data layouts are different. The actual data included on the invoices can vary. The placement of key data fields depends on the type of invoice and the vendor that issues it. A sophisticated OCR solution should be able to identify these variations, without requiring invoice “zones”, templates or vendor definitions.

That’s why we built ktIQ.

ktIQ is an intelligent, AI-modeled data extraction service that is powered by real-time integration with the ERP system, in this case – Dynamics GP.

Here are three key differences between traditional OCR and ktIQ.

Traditional (Zonal) OCR ktIQ – Intelligent Data Extraction
·     Extensive, expensive setup and ongoing maintenance of invoice/document templates ·     Significantly reduced setup time and costs since invoices are read like a human would, using artificial intelligence
·     Requires user interaction to improve data accuracy and character confidence ·     Automatically determines invoice type based on the data it reads and verifies
·     Many vendor invoices need unique definitions and document zones defined ·     Improves data accuracy automatically without user interaction


Now, to be clear, the initial release of ktIQ is designed specifically for KwikPayables with Dynamics GP integration. It is the front-end, add-on module that performs the automatic data extraction and invoice indexing step in the process.

Watch the video below to see ktIQ in action with a non-P.O. invoice workflow.

If you would like to see ktIQ in action with a P.O. invoice workflow, please reach out to Michele Lewis (mlewis@enchoice.com) to discuss how ktIQ would work for your organization.

Check back often for updates. As ktIQ gets smarter and extends its abilities, we’ll share it with you here!