How Amazon Connect Health brings agentic AI to the point of care

This article introduces Amazon Connect Health, an agentic AI service that integrates clinical documentation, patient insights, and medical coding directly into existing EHR workflows to reduce clinician administrative burden.

• Patient insights capability generates pre-visit summaries from HealthLake FHIR data with evidence-linked references
• Ambient documentation captures real-time patient-clinician conversations and generates SOAP notes, medical codes, and after-visit summaries
• Standalone medical coding API accepts clinical text and returns ICD-10 and CPT codes with confidence scores
• Single unified SDK with consistent interface across all three capabilities and organizational domain model
• Supports multiple programming languages: Java, Python, JavaScript, .NET, Go, C++, Ruby, Rust, Swift
• All outputs written to Amazon S3; runs on HIPAA-eligible AWS infrastructure with IAM-based security
• Available in US East (N. Virginia) and US West (Oregon) regions
• Sample EHR integration application available on GitHub with CloudFormation deployment templates

Amazon Connect Health embeds AI capabilities directly into clinician workflows rather than requiring separate applications, enabling healthcare builders to reduce documentation time and clinician burnout through modular, standards-based integration.