Insurance Claim Management Agent

Agent Description:

The Insurance Claim Management Agent streamlines the healthcare insurance claim lifecycle by automating data retrieval and consolidation. It follows a two-step process: first, the Claim Data Extractor fetches patient visit and diagnosis data via a GET tool and structures it into JSON. Second, the Claim Consolidation & Analysis Agent queries an SQL database using the Claim Reference ID to retrieve related insurance policy, coverage, and claim settlement details. Finally, it merges all this information into a single, unified summary. This automated consolidation of patient and policy data provides a complete, structured view of the claim, significantly accelerating the review and processing workflow.

Purpose and Components
  • Purpose: To streamline the healthcare insurance claim lifecycle by automatically fetching patient visit details and insurance policy information, merging them based on a claim reference, and providing a unified summary for faster review and processing.
  • Components:

    • Claim Data Extractor: An agent to fetch and structure patient visit and diagnosis data from an API source.

    • Claim Consolidator & Analyzer: An agent to query an insurance database for policy details and merge this with the patient data.

    • Tool Integrations: Pre-configured tools for making GET requests and interacting with an SQL (SQLite) database.

Supported Capabilities

  • Retrieving JSON reports containing patient details and visit summaries via a GET tool.

  • Parsing and extracting key entities like Patient ID, Name, Age, Gender, Diagnosis, Treatment, Date of Visit, and Claim Reference ID.

  • Converting extracted content into a structured JSON format.

  • Querying an insurance SQL database using a Claim Reference ID.

  • Retrieving insurance details such as Policy Number, Insurance Provider, Coverage Type, Claim Status, Approved Amount, and Remarks.

  • Merging patient visit data (from API) and insurance data (from SQL) into a single, unified output object.

  • Returning a structured summary that includes patient details, insurance coverage information, and the current claim settlement status.

LLM Used

  • GPT_4O_MINI

    Note: To learn more about the LLM and to modify its behavior, refer to the Configuring LLM settings section.

Sub-Agents

1. Claim Data Extractor

  • Role: Claim Data Extractor

  • Scope: Identify patient details, treatment codes, billing amounts, and hospital/provider info from unstructured claim forms.

  • Description: This agent initiates the process by using the GET tool to retrieve a JSON report containing patient details and visit summaries. It parses this data, extracts key entities (like Patient ID, Diagnosis, Claim Reference ID), and converts it into a structured JSON format, ready for the next step.

  • LLM Used: Default (Inherits from parent agent).

  • Tools Used: Request - Get

2. Claim Consolidation and Analysis

  • Role: Insight generator

  • Scope: Consolidate all extracted claims into a central claim ledger and highlight potential duplicates, errors, or unusual costs.

  • Description: This agent receives the structured patient JSON from the first agent. It then uses the SQL toolkit to query the insurance database, using the Claim Reference ID to fetch corresponding policy and claim status details (like Policy Number, Approved Amount). Finally, it merges the patient data and insurance data into one unified object and outputs a structured summary.

  • LLM Used: Default (Inherits from parent agent).

  • Tools Used: SQL - Toolkit

Tools Used:

  • Request - Get Tool: Fetches the patient visit and claim details from a remote JSON endpoint.

  • SQL - Toolkit (SQLite): Executes queries against a SQLite database to retrieve insurance policy and claim status information based on the Claim Reference ID.

Note: For details on modifying the Tools, refer Tools Library section.
Agent Workflow Behavior Summary

  1. The Claim Data Extractor (start node) uses the Request - Get tool to fetch patient and visit data for a specific claim.

  2. It structures this information into a JSON object, including the Claim Reference ID.

  3. This structured patient JSON is passed to the Claim Consolidation & Analysis agent.

  4. The Claim Consolidation & Analysis agent (end node) uses the Claim Reference ID and the SQL - Toolkit to query the insurance database for policy and claim status details.

  5. It merges the patient data (from step 2) and the insurance data (from step 4) into a single, unified output object, providing a complete summary of the claim.

Sample Questions:

  • Process insurance claim with Reference ID 'CR1005'.

  • Fetch patient details and insurance status for Claim 'CR1006'.