Automobile Insurance Claim Processing with Amazon Bedrock > Introduction to Automobile Insurance Claim Processing with Amazon Bedrock recipe > Automobile Insurance Claim Processing with Amazon Bedrock recipe contents
  

Automobile Insurance Claim Processing with Amazon Bedrock recipe contents

The recipe contains multiple assets, such as a process objects, app connections, human task, processes, and guide.
The following image shows the assets that the Automobile Insurance Claim Processing with Amazon Bedrock recipe package contains:
The image shows the assets that the Automobile Insurance Claim Processing with Amazon Bedrock recipe package contains.

Automobile Insurance Claim Processing with Amazon Bedrock recipe assets

The following table lists the assets that the Automobile Insurance Claim Processing with Amazon Bedrock recipe package contains:
Asset Name
Asset Type
Description
VehicleDetails
Process object
Provides the required vehicle details to process the insurance claim.
ClaimRequest
Process object
Provides the required fields to process the insurance claim.
AmazonBedrockConnection
App connection
Amazon connection that connects to Amazon Bedrock and performs actions.
EmailConnection
App connection
Email connection to send an email.
Check Image Format
Process
Subprocess that verifies the file type.
Get Price List and Parts List
Process
Subprocess to get a sample price list and a list of parts for a specified vehicle.
Send Email with Claim Details
Process
Subprocess that sends an email with the claim details.
Calculating the Payout
Process
Subprocess that calculates the payout based on damaged parts and their prices for a specified vehicle.
Validate Vehicle Information
Process
Subprocess that checks vehicle details from a sample vehicle data set.
Damage Recognition using Amazon Bedrock LLM
Process
Subprocess that calls an LLM model to determine a list of damaged parts.
Process Claim Request
Process
The process is called by an HTTP request with basic information about the insurance claim request as an incoming parameter. The process generates a claim ID and then performs initial vehicle verification by validating information from the data set. After successful validation, the process gets a sample price list according to the chosen vehicle. The process then validates the image format and performs damage recognition using the specified LLM model. If the damage level is simple, the process performs calculations for an approximate payout for the compensation. If the damage level is complex, the application is sent for review to determine the list of damaged parts associated with the claim based on the applicant's uploaded images. Based on the reviewer's decision, the process also performs calculations for the approximate payout for the compensation and sends a message to the specified email address with the approximate compensation price.
DamageDetermination
Human task
Determines the damaged parts based on the image of the incident.
Insurance Claim Processor
Guide
Demonstrates insurance claim processing with GenAI.