Devron | Financial Document Categorization

Financial Document Categorization

A financial institution needed to isolate relevant intelligence from various data types amid a sea of information.

Inefficiencies in document verification to classify bank statements, paystubs, and tax forms for loan approvals and risk underwriting.
Data dispersed across different mortgage data applications often hindering the ability to quickly scan documents of interest.
Lack of data science depth to train highly efficient algorithms on dispersed datasets limited the accuracy of document categorization.

Devron deployed across multiple loan data applications that housed sensitive documents (e.g., bank statements, tax-ids, W-2s) instead of duplicating and centralizing data.
Devron Central Authority trains algorithms to understand each type of document using computer vision and high accuracy image classification techniques.
With more training data and time saved from unnecessary centralization, Devron predicts loan approval and risk underwriting documents of interest with over 95% accuracy.

Financial Document Categorization