Administrative Agent

Agent Description:

The Administrative Agent automates healthcare administration by processing patient requests from an API. It identifies tasks such as scheduling or claim verification, uses an SQL database for data handling, and returns accurate summaries or confirmations - improving overall administrative efficiency.

Purpose and Components
  • Purpose: To automate the handling of patient administrative requests, from initial API contact to database query and resolution, ensuring timely and accurate processing of scheduling, claims, and queries.
  • Components:

    • API Request Handler: An agent to ingest and understand patient requests received via a GET API.

    • Database Agent: An agent to query, update, and retrieve information from patient-related SQL databases.

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

Supported Capabilities

  • Receiving patient requests via a GET API.

  • Extracting key fields: patient_id, request_type, requested_time, query, claims_doc_submitted.

  • Normalizing and structuring incoming request data.

  • Identifying request types (new appointment, follow-up, cancellation, or claims inquiry).

  • Querying SQL tables (PatientAppointments, PatientClaims) using patient details.

  • Retrieving current appointment status, upcoming schedules, and claim document information.

  • Updating records for new appointments, cancellations, or follow-ups.

  • Generating a concise summary of actions taken, next steps, and confirmations.

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. GET Request Handler

  • Role: Request Collector

  • Scope: Receive and process patient requests via GET API

  • Description: This sub-agent serves as the front-end, receiving patient requests via a GET API. It extracts key fields (like patient_id, request_type), normalizes the data, identifies the request type (e.g., new appointment, claims inquiry), and then forwards a structured data package to the Database Response Agent for processing.

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

  • Tools Used: Request - Get

2. Database Response Agent

  • Role: Data Resolver

  • Scope: SQL Database (PatientAppointments, PatientClaims)

  • Description: This sub-agent interacts directly with the patient database. It queries the PatientAppointments and PatientClaims tables based on the request details. It retrieves current statuses, updates records (for new appointments or cancellations), generates a concise summary of actions taken and next steps, and prepares the final response.

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

  • Tools Used: SQL - Toolkit

Tools Used:

  • Request - Get Tool- Receives patient data via GET API requests from a specified JSON endpoint.

  • SQL - Toolkit (SQLite)- Executes queries, updates records, and retrieves schema information from the patient SQLite database.

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

  1. The GET Request Handler (start node) receives a patient request via its GET API tool.

  2. It processes the request, extracts key data (like patient_id and request_type), and identifies the task.

  3. It then forwards the structured request data to the Database Response Agent.

  4. The Database Response Agent uses its SQL Toolkit to query and/or update the patient database (e.g., PatientAppointments, PatientClaims).

  5. It generates a final summary, confirmation, or status update based on the database transaction.

  6. The Database Response Agent (end node) provides the final structured output.

Sample Questions:

  • I need to schedule a new appointment for patient_id 'P1001' for 2023-11-15.

  • What is the status of the claim for patient_id 'P1002'?