AI-Powered Document Classification

AI-Powered Document Classification

When you upload documents, ABDP automatically identifies what type each document is. The system has been trained to recognize common loan document types like:

● Credit Applications

● Lease Agreements

● Security Agreements

● Amortization Schedules

● Invoices

● Insurance Certificates

● Title Documents

● UCC/PPSA Filings


Confidence Indicators:Each classification includes a confidence indicator:

● 🟢 Green - High confidence. The AI is very sure about this classification.

● 🟠 Orange - Medium confidence. You should verify this is correct.

● 🔴 Red - Low confidence. Manual review required.


    • Related Articles

    • Configuring Document Types

      For each queue, specify which document types you expect to receive: 1. Go to Queue Settings > Document Types 2. Add document types (e.g., Lease Agreement, Credit Application) 3. Mark each as Required or Optional 4. The AI will be configured to ...

    • ABDP Value Proposition

      Before ABDP: ● Hours spent manually reviewing each loan file ● High risk of human error in data entry ● Difficult to scale during busy periods ● Paper-based audit trails ● Inconsistent validation processes With ABDP: ● Documents processed in minutes, ...

    • Quality Metrics

      Track accuracy and issues: ● Classification Accuracy - How often AI correctly identifies document types ● Extraction Accuracy - How often extracted values are accepted without correction ● Validation Pass Rate - Percentage of loans that pass all ...

    • Reviewing Classifications

      After processing completes, review how documents were classified: 1. Open your batch 2. Click on any folder (loan) 3. Review the documents and their assigned types 4. Look for orange or red confidence flags If a document is misclassified: 1. Click on ...

    • Analytics & Reporting

      ABDP tracks everything and provides dashboards to help you understand your processing performance: What You Can See: ● How many batches, folders, and documents you've processed ● Processing times (average, fastest, slowest) ● Classification accuracy ...