Ask Infrrd #2: Why Are There So Few Successful Mortgage OCR Solutions?

 


Welcome to the second post in the Ask Infrrd series - your platform to have an honest, candid conversation about everything related to document data extraction. In this series we answer frequently-asked questions for our IDP experts or that we receive at hello@infrrd.ai. Here, we answer another question about Mortgage processing.

Q: Why are many mortgage companies unable to find an effective OCR partner who can show results?

A: Before we delve into the answer, let us take a step back and learn more about why OCR solutions have struggled to deliver effectiveness for most industries, not just Mortgage. OCR technology was designed to solve a different problem than data extracting. OCR was a means to convert images with text to digitized text that can be read on a computer. If we are only considering this function of OCR, then I would argue that it has done a phenomenal job and come a long way since it was originally created. OCR engines of today can read very low-quality images with very high accuracy.

The disconnect is that customers do not just need the digitized text from OCR. They expect technology to understand the digitized text and answer complex questions, such as:

What is the employer's name on this W2?
What was the account balance at the end of the month?

OCR technology was not designed for this. That is the job of Intelligent Document Processing(IDP). IDP is an ecosystem combining Natural Language Processing, Computer Vision, and Predictive Analytics, all rolled into one comprehensive solution for understanding documents and answering questions. Machine learning and AI capabilities filled the gaps OCR was not designed to address.

IDP solutions today look at numerous signals fed into several machine learning algorithms to read and understand documents. Thus, IDP has the capabilities to extract data that could not be extracted by OCR. It also enables the ability to handle millions of variations of documents without relying on templates.

This is why an effective OCR solution has eluded most industries and IDP solutions can provide mortgage providers an effective solution to understand and process documents.

We will continue this conversation in our next post where we will talk about the complexity of title/purchase agreements and contracts to signature detection. The key takeaway from this first conversation is the understanding that IDP solutions are different from OCR engines. IDP platforms can recognize the information’s context, structure, and presentation. Even if OCR systems have let you down in the past, have faith and give IDP a spin. Our mortgage customers have been pleasantly surprised by what IDP can do and we would love to convert your skepticism into belief.

Comments

Popular posts from this blog

What Mick Jagger Taught Me about Data Extraction from Tables

Eliminate your OCR and Manual Data Entry Bottleneck

How to Prepare Data For OCR Learning