Overview
Auto-approval allows eligible claims to be automatically approved and resolved by the AI without manual review.
This helps:
Reduce support workload
Speed up customer resolution
Maintain consistent and policy-driven decision-making
When a claim meets all configured criteria, the system will automatically approve and process the resolution.
How Auto-Approval Works
When a claim is submitted, the system evaluates it against a set of rules defined in your delivery policies.
If all conditions are met, the claim is automatically approved
If any condition fails, the claim is routed to manual review
Auto-Approval Logic
A claim is auto-approved only if it passes all of the following checks:
Within Value Threshold
The claim value must be below the configured auto-approval limitWithin Filing Window
The claim must fall within the defined lower and upper filing thresholdsFraud Check
The claim must pass fraud validation, including:Being the first claim submitted by the customer/email
Reason-Specific Checks
Additional validations based on the claim type
Reason-Specific Checks
Certain claim types require additional verification before they can be auto-approved:
Claim Reason | Required Check |
|---|---|
Damaged | Courier marked as damaged OR AI image evaluation confirms damage |
Defective | AI image evaluation |
Lost in Transit | Courier marked as lost |
Misdelivered | Mismatch between delivery and shipping address |
Wrong Item | AI image evaluation |
For all other claim reasons, the system defaults to approval if other conditions are met.
AI Image Evaluation
For applicable claim types (e.g., damaged, defective, wrong item):
The AI analyzes uploaded images
Determines whether the claim is valid based on visual evidence
Reduces the need for manual review while maintaining accuracy
When Claims Are Not Auto-Approved
A claim will be routed to manual review if any of the following conditions are met:
It exceeds the auto-approval value threshold
It falls outside the filing window
Required evidence is missing
It fails fraud checks
The customer has previously submitted a claim (not the first claim for that email/customer)
It does not pass reason-specific validation
Best Practices
Start with conservative auto-approval thresholds
Require uploads for higher-risk claim types
Enable AI image detection where applicable
Monitor claim outcomes and adjust rules over time
Key Takeaway
Auto-approval enables fast, automated claim resolution while still enforcing:
Policy rules
Fraud safeguards
Evidence validation
Only eligible and verified claims are automatically approved, ensuring both efficiency and control.
If you need any assistance, please contact support@cavalry.ai.