OCR – PROSAR-AIDA for Mortgage Loans

Title. 4506-T. Deed.

Convert your unstructured images into indexed searchable PDF and into structured data.

PROSAR-AIDA is text recognition and analysis software which can determine document types, classify images, and capture or extract data for routing and archival.

A Mortgage solution requires PROSAR-AIDA’s unique ability to process an unlimited number of versions and formats of particular document types. This capability is unique to PROSAR-AIDA and a vital feature for creating a true and effective unstructured document classification and data capture solution.

  • With other solutions the document content is expected and assumed to always appear in the same geographical areas on a document.  Our solution functions more like a human, reading all content regardless of location.
  • The process begins with a high-speed (sub-second) and efficient full-page OCR scan of each image.
  • This first step allows PROSAR-AIDA to search each and every word of every document and discover the content much as a human being would.
  • This overall process of discovery is unique, and makes PROSAR-AIDA very different from the typical structured or template-based solution.

Validation – PROKEY for Mortgage Loans

PROKEY is an image-assisted exception processing application which provides a highly intuitive user interface for managing the quality assurance process.
  • Image Review
  • Metadata review/verification/entry
  • Document Boundary View

PROKEY is tightly integrated with our advanced OCR application, PROSAR-AIDA, and offers an ability to create very complex database connections and business rule logic checks in order to assure extremely accurate output.
An operator is presented with uncertain page types or data and then prompted to accept or modify what is presented by the system.
Some of the useful features of PROKEY include the ability to:

  • Merge/break documents’ boundaries
  • Prompt the operator with “most likely alternatives” if the primary choice from the recognition engine is not chosen.
  • Use multiple monitors
  • Present data elements for validation with image snippets nearby for key entry efficiency
  • Allow for customized validation error and warning messages related to database or value range validations.

Bookmarked PDF Output Module

Mortgage Loan documents and full Loan packages, can be exported as PDF documents which include a full Table of Contents.

Document types within a full loan package of 500 or more pages, can be quickly referenced using a standard PDF viewer and a couple clicks of a mouse.

Mortgage Loan document data elements can be quickly identified with the click of a mouse within a PDF viewer.

Redaction Module

  • Paradatec’s OCR technology is able to read an entire page of an 8.5” x 11“ document in less than 500 milliseconds on average. It then allows a rules-based search engine to locate sensitive information contained within the document also in milliseconds.
  • Automatically finds and removes personally identifiable information (“PII”) in sensitive documents.
  • Makes data protection a seamless part of everyday workflows, not a separate manual process
  • Streamlines compliance with legal requirements for privacy and disclosure
  • Protects PII from unintentional disclosure, theft, or other misuse
  • Replaces tedious, error prone, manual redaction with a highly efficient automated process
Paradatec Auto Redaction Solution, unredacted example
Paradatec Auto Redaction Solution, after less than 1 second discovery and redaction process:

WriteUCD - Paradatec's UCD Creation Module

The new Closing Disclosure documents, and the regulations related to them, put a new emphasis on the need for quick and very accurate data.

For all newly originated loans with a Note Date of September 25, 2017 or later, the GSE lenders (Fannie Mae, and Freddie Mac) will require that a Uniform Closing Dataset (“UCD”) file and a PDF copy of the related buyer-side and seller-side Closing Disclosures be included for any loan they acquire.

  • In less than a second each page of a Closing Disclosure document is fully read and hundreds of data fields are captured.
  • The end result is 75-80% of the work is completed more accurately than with human key entry and in a tiny fraction of the time.

Paradatec’s new WriteUCD solution further extends these capabilities by formatting efficiently captured Closing Data according to the MISMO 3.3 UCD standard from both a buyer-side and seller-side Closing Disclosure and federating the data into a consolidated UCD file.