AWS Machine Learning Blog
In this post, we discuss how generative artificial intelligence (AI) can help health insurance plan members get the information they need. The solution presented in this post not only enhances the member experience by providing a more intuitive and user-friendly interface, but also has the potential to reduce call volumes and operational costs for healthcare payers and plans. {text2} </user_question> The {text1} and {text2} data items will be replaced programmatically to populate the ID of the logged-in member and user question. Also, more examples can be added to help the LLM generate appropriate SQLs. The following is an example of a data summarization prompt template: <role> You are a customer service agent working for a health insurance plan and helping to answer questions asked by a customer. </role> <task> Use the result_dataset containing healthcare claims data to answer the user_question. This result_dataset is the output of the sql_query. </task> <instructions> 1. To answer a question, use simple non-technical language, just like a customer service agent talking to a 65-year-old customer. 2. Use a conversational style to answer the question precisely. 3. If the JSON contains a "count" field, it means the count of claims. For example, "count": 6 means there are 6 claims, and "count": 11 means there are 11 claims. 4. If the result_dataset does not contain meaningful claims data, then respond with one line only: "No data found for the search criteria." </instructions> <user_question> {text1} </user_question> <sql_query> {text2} </sql_query> <result_dataset> {text3} </result_dataset> The {text1}, {text2}, and {text3} data items will be replaced programmatically to populate the user question, the SQL query generated in the previous step, and data formatted in JSON and retrieved from Amazon RDS. Security Amazon Bedrock is in scope for common compliance standards such as Service and Organization Control (SOC), International Organization for Standardization (ISO), and Health Insurance Portability and Accountability Act (HIPAA) eligibility, and you can use Amazon Bedrock in compliance with the General Data Protection Regulation (GDPR). The service enables you to deploy and use LLMs in a secured and controlled environment. The Amazon Bedrock VPC endpoints powered by AWS PrivateLink allow you to establish a private connection between the virtual private cloud (VPC) in your account and the Amazon Bedrock service account. It enables VPC instances to communicate with service resources without the need for public IP addresses. We define the different accounts as follows: Customer account This is the account owned by the customer, where they manage their AWS resources such as RDS instances and Lambda functions, and interact with the Amazon Bedrock hosted LLMs securely using Amazon Bedrock VPC endpoints. You should manage access to Amazon RDS resources and databases by following the security best practices for Amazon RDS. Amazon Bedrock service accounts This set of accounts is owned and operated by the Amazon Bedrock service team, which hosts the various service APIs and related service infrastructure. Model deployment accounts The LLMs offered by various vendors are hosted and operated by AWS in separate accounts dedicated for model deployment. Amazon Bedrock maintains complete control and ownership of model deployment accounts, making sure no LLM vendor has access to these accounts. When a customer interacts with Amazon Bedrock, their requests are routed through a secured network connection to the Amazon Bedrock service account. Amazon Bedrock then determines which model deployment account hosts the LLM model requested by the customer, finds the corresponding endpoint, and routes the request securely to the model endpoint hosted in that account. The LLM models are used for inference tasks, such as generating text or answering questions. No customer data is stored within Amazon Bedrock accounts, nor is it ever shared with LLM providers or used for tuning the models. Communications and data transfers occur over private network connections using TLS 1.2+, minimizing the risk of data exposure or unauthorized access. By implementing this multi-account architecture and private connectivity, Amazon Bedrock provides a secure environment, making sure customer data remains isolated and secure within the customer’s own account, while still allowing them to use the power of LLMs provided by third-party providers. Conclusion Empowering health insurance plan members with generative AI technology can revolutionize the way they interact with their insurance plans and access essential information. By integrating conversational AI assistants powered by Amazon Bedrock and using purpose-built AWS data services such as Amazon RDS, healthcare payers and insurance plans can provide a seamless, intuitive experience for their members. This solution not only enhances member satisfaction, but can also reduce operational costs by streamlining customer service operations. Embracing innovative technologies like generative AI becomes crucial for organizations to stay competitive and deliver exceptional member experiences. To learn more about how generative AI can accelerate health innovations and improve patient experiences, refer to Payors on AWS and Transforming Patient Care: Generative AI Innovations in Healthcare and Life Sciences (Part 1). For more information about using generative AI with AWS services, refer to Build generative AI applications with Amazon Aurora and Knowledge Bases for Amazon Bedrock and the Generative AI category on the AWS Database Blog. About the Authors Sachin Jain is a Senior Solutions Architect at Amazon Web Services (AWS) with focus on helping Healthcare and Life-Sciences customers in their cloud journey. He has over 20 years of experience in technology, healthcare and engineering space. Sanjoy Thanneer is a Sr. Technical Account Manager with AWS based out of New York. He has over 20 years of experience working in Database and Analytics Domains. He is passionate about helping enterprise customers build scalable , resilient and cost efficient Applications. Sukhomoy Basak is a Sr. Solutions Architect at Amazon Web Services, with a passion for Data, Analytics, and GenAI solutions. Sukhomoy works with enterprise customers to help them architect, build, and scale applications to achieve their business outcomes.
Published: 2024-09-12T19:13:52
