Staircase, the company building an integrated, digital infrastructure to accelerate tech-enabled mortgages, has introduced a new software that helps business process outsourcers (BPOs) reduce labor costs by improving quality control in the mortgage document verification process.
Many lenders are outsourcing certain functions to lower costs and overcome the challenges of operating remote teams. One of the main tasks outsourced to BPOs is document verification which relies heavily on manual labor and off-the-shelf optical character recognition (OCR) technology to read documents. Staircase’s Quality Control automates the document verification process by seamlessly combining a BPO’s existing OCR provider with labeling software in a tight training loop where the labeling software guesses the value of the data missed by the OCR engine, the BPO staff verify the correctness of the data, and this new ground truth feeds back into the OCR engine to improve accuracy on a daily basis. The improved extraction accuracy leads to fewer and fewer documents being sent to the manual exception handling process, saving enormous time and cost.
Recommended AI News: Data Sprawl Creating Risk for Organizations Worldwide as Personal App Use in Business Continues to Rise
“Document verification solutions that are efficient and affordable are critical to the mortgage process,” said Staircase CEO and cofounder Adam Kalamchi. “Our Quality Control product helps third parties reduce labor costs by enabling OCR to get better and better over time with a self-learning loop. We of course also include advanced data and document security to ensure privacy standards are met.”
“Staircase’s new technology monitors and improves OCR results daily, resulting in increased cost savings over time,” said Edmond Cakulla, Staircase product engineer. “It also tracks the productivity of human labelers which can be used for advanced training for current and new users.”
With Staircase’s technology, OCR providers can be easily changed in or out of the document verification process or can even be combined. The technology can be implemented in as little as one week.
Recommended AI News: Connectbase to Power Nctc’s Connectivity Exchange Platform
[To share your insights with us, please write to sghosh@martechseries.com]
Comments are closed.