Over the last few years of working with my business partner on designing an effective income analysis software tool, we have been asked this question on numerous occasions.
“Does your software use OCR to give income automatically?”
We both agree it was a fair question and thought it would be a great idea to build into our design. However, after hours of research and programming we found some startling answers about OCR for income review that we thought worth sharing.
Issues with OCR
First, OCR only works for just a few of the 30+ income types listed on each of the agency Allregs sites. We found that creating an income tool that only worked on a minority number of loans was not a good solution.
Second, for the few income types it does work on it still requires a trained underwriter to review the tax returns to compare data for accuracy and agency guideline cash flow adjustments. For example if line 6 on a schedule C (a non-recurring income field) has $5,000 listed, would the OCR know if it should ignore or keep that number. With the time needed to gather the documents, upload them to the OCR service, and review the OCR form for accuracy, he or she could have entered the data into UberWriter and be onto the next file. This is results in little or no efficiency gains for the income review.
Third, since OCR just puts numbers on forms it provides no guidance, training, or safety checks on the income reviewed. These leaves many companies in the same situation of only a very few trained team members the ability to review the forms.
Why we don’t advocate OCR for underwriting
These are just some of the concepts that we felt did not provide the capability we wanted to provide our clients. In next week’s blog I will provide some more detail on each point so that others who are taking on the same challenge have some notes to compare too.