The Uniform Residential Appraisal Report document skill for ABBYY Vantage is an AI model that extracts data from US and Canadian uniform residential appraisal reports (Fannie Mae Form 1004, Freddie Mac Form 70), which are documents used to provide information about residential real estate and its estimated value. These documents are also required for mortgage loan applications and contain the following information: information about the property and its surrounding area, real estate market data, technical data, property photos, property blueprints, etc.
The Uniform Residential Appraisal Report (Appraisal) skill is a preview skill. It has been trained on a small document set and is intended to be used by citizen developers as a quick-start basis for processing appraisals. The skill may require further training on your specific documents in order to be used in production scenarios.
Benefits
- Vantage document skills are extraction models, based on ABBYY's purpose-built AI, designed and trained to extract key data from business documents.
- Preview document skills provide a quick-start basis for conducting proof of concept and automating document-centric processes, without wasting time to train a new skill from the ground up. In most preview skills the main data fields have been labeled and the skill has been trained on a small document set to provide a starting point for further training on customer side. Preview skills generally come at no additional cost, but consume page volume from the base Vantage Core Skills volume.
- The AI model automatically detects key fields on Fannie Mae Form 1004 and Freddie Mac Form 70 and extracts relevant data (file number, property address, borrower, neighborhood details, improvements etc.). For a full list of extracted fields, please see the skill documentation.
- Business rules for data validation can be added as required.
- Can be augmented by human-in-the-loop review and verification based on accuracy thresholds, missing or incorrect information.
- Requires further training in order to ensure sufficient accuracy levels on customer-specific documents. Training can be done up-front or during runtime as documents arrive and go through manual verification.