How it Works / How to Build It

1. Claim Intake: The ClaimIntake node processes and validates initial claim information, creating a structured summary from various inputs like policy number, claimant name, and supporting documents.
2. Risk Assessment: The RiskAssessor node evaluates the claim for potential fraud and validity, using indicators such as claim timing and documentation quality.
3. Policy Verification: The PolicyVerifier node checks the policy details against the claim to ensure coverage and eligibility.
4. Damage Assessment: The DamageAssessor node assesses the extent of damage and validates repair estimates based on visual evidence.
5. Document Analysis: The DocumentAnalyzer node extracts key information from claim-related documents to ensure consistency and compliance.
6. External Data Verification: The ExternalDataVerificationAgent node verifies claim information against external sources, such as DMV records and fraud databases.
7. Adjudication: The Adjudicator node makes the final decision based on all assessment inputs, providing a structured decision rationale.
8. Final Outputs: The workflow generates outputs for processing status, risk score, audit trail, and the final adjudication decision using nodes like FinalOutputProcessingStatus, FinalOutputRiskScore, FinalOutputAuditTrail, and FinalOutputDecision.

What You Can Use This For

• Insurance claims processing
• Fraud detection and risk assessment
• Policy compliance verification
• Damage evaluation and repair cost validation
• Generating audit trails for regulatory compliance

Prerequisites

• Vellum account
• Access to relevant claim documents (e.g., policy documents, medical records)
• Input data for claims (e.g., claimant name, claim amount, incident date)

How to Set It Up

1. Clone the workflow template in your Vellum account.
2. Configure the Inputs class with your specific claim data.
3. Ensure all necessary documents are uploaded and accessible.
4. Connect the nodes as outlined in the workflow graph.
5. Test the workflow with sample claims to validate functionality.
6. Customize any prompts or templates as needed for your specific use case.