AWS Builder Center

Automate CMS-1500 Form Filing with Strands Agent and Amazon Bedrock

An AI-powered insurance data processing system that transforms patient insurance information from JSON format to structured CSV output using Amazon Bedrock and the Strands Agents SDK. This tool automates the complex task of mapping insurance data fields according to predefined schemas, making it ideal for healthcare organizations processing patient coverage information.

Healthcare providers often spend significant time manually completing CMS-1500 forms, which capture essential patient and insurance information for claims processing. This manual process is time-consuming, error-prone, and delays reimbursements. In this post, we demonstrate how to automate CMS-1500 form filing of demographic and coverage information with AI agent by leveraging  Amazon Bedrock ,  Amazon SageMaker , and the  Strands Agents SDK . This AI agent reduces the manual time to interpret and enter this information from several hours in a day to seconds. We'll show you how to implement this solution to dramatically reduce administrative overhead and allow healthcare staff to focus more time on patient care.

Key Challenges in Filling CMS-1500 Form

The CMS-1500 form, also known as the HCFA (Health Care Financing Administration) 1500 Form, is a standardized form used by healthcare providers to bill insurance companies for services rendered, primarily for outpatient care. It's commonly used to submit claims to  Medicare ,  Medicaid , and other government and private insurance programs. Traditionally, form completion requires meticulous manual entry to ensure compliance with payer requirements.
The key challenges in filing CMS-1500 Form are:
  • Manual Effort: Each claim involves re-entering structured data from disparate sources.
  • Time-Consuming: Processing delays can stretch over days or weeks.
  • Error-Prone: Manual entry increases the risk of claim rejections due to inconsistencies.

How we solve this problem?

We solve the problem by creating a golden schema for the CMS-1500 form and automating patient data mapping with AI Agent which eliminate repetitive manual steps. Using Amazon Bedrock’s foundation models and the Strands Agents SDK, we transform raw JSON patient data into CMS-1500-compatible CSV or JSON output, ready for claim submission.

Solution Architecture

The below diagram depicts the architecture of the solution:


Technical Steps

Prerequisites
  • Python: Version 3.10 or higher
  • AWS Account: Active AWS account
  • Amazon Bedrock Access: Access to Amazon Bedrock service in us-east-1 region for the following models:
  • Claude 3.5 Sonnet v2: use cross region inference | us.anthropic.claude-3-5-sonnet-20240620-v1:0 model
  • Nova Pro: use cross region inference | us.amazon.nova-pro-v1:0modell
  • llama4-scout-17b-instructt: use cross region inference | us.meta.llama4-scout-17b-instruct-v1:0
  • AWS Credentials: Properly configured AWS credentials (via AWS CLI, IAM roles, or environment variables)
  • Amazon SageMaker Studio: Access to Jupyter notebook for development (optional)
Installation
Clone the repository:  git clone https://github.com/aws-samples/sample-strand-agents-CMS-1500-form.git 
Follow the instructions of the GitHub  link 

Benefits

  • Time Savings: Reduces processing time from days to minutes.
  • Increased Accuracy: AI-driven mapping minimizes human error.
  • Scalability: Supports multi-payer claims with primary, secondary, and tertiary insurance mapping.
  • Operational Efficiency: Frees staff to focus on higher-value tasks.

Conclusion

This solution shows how Amazon Bedrock, Amazon SageMaker, and Strands Agents can automate a traditionally manual healthcare process to bring efficiency. By leveraging Agentic AI patterns, you can streamline insurance claim preparation, reduce operational costs, and improve turnaround times. Start building your own Agentic AI solutions with  Amazon Bedrock  and  Strands Agents  to transform repetitive workflows into intelligent, automated pipelines.

Author:

Purna Sanyal, Generative AI Specialist Architect at AWS. He has specialization in data strategy, machine learning, Generative AI, Agentic AI and cloud native architecture.
George Krucik, BSc (CS) MD MBA is currently the founder and CTO of iTreatMD an ONC Certified EMR delivering AI driven applications that provide practicing physicians, surgeons and nurse practitioners with tools to expedite their point of care documentation needs.
Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article