Amazon Web Services is moving deeper into agent-based artificial intelligence for healthcare, targeting common administrative tasks such as scheduling, medical history, clinical documentation, and coding.
Amazon Connect Health, an agent-based AI solution, is the first AWS solution specifically for healthcare providers, the company announced Thursday. Developed for healthcare providers and health technology companies, Amazon Connect Health provides five agent AI capabilities that use natural language voice technology to support administrative tasks such as patient identification and appointment scheduling for call center staff, while reducing manual work for clinicians by synthesizing patient medical histories and leveraging ambient technology to transcribe consultations. The solution also streamlines post-visit medical coding for patients and providers.
“In conversations with large health systems, AWS found that staff spend up to 80% of call time manually editing data across fragmented tools: verifying patient identities, scheduling appointments, manually piecing together patient histories spread across multiple systems, and meeting complex documentation requirements. These tasks take clinicians and their teams away from patient care. 89% of patients “Care navigation challenges, such as scheduling difficulties, long wait times, and access barriers, are the reason for the switch,” Colleen Aubrey, senior vice president of applied AI solutions at AWS, said in a blog post.
The company says its AWS solution can be deployed in days instead of months with native electronic health records and Amazon Connect contact center integration, pre-built connections to over 100 EHRs through data integrator partners, and a fully managed integration SDK.
If a patient calls their health care provider and says, “I want to see my doctor next week after work.” Amazon Connect Health understands the reason for the call, the patient’s situation, and preferences, Aubrey said. “Verify who the patient is, verify insurance, check patient and provider availability, and schedule an appointment while the patient is on the phone. Amazon Connect Health combines the power of Connect, AWS’s AI-powered customer experience solution, with real-time connectivity to your EHR,” Aubrey wrote.
For situations that require a human response, such as medical concerns or complex requests, solutions are seamlessly escalated to staff. According to AWS, health systems can customize exactly when and how handoffs occur.
Some health systems have already deployed the capabilities of Amazon Connect Health. According to a blog post, UC San Diego Health, which handles 3.2 million patient interactions annually, reports saving one minute per call, redirecting 630 hours each week from patient identification to direct patient assistance, and reducing call abandonment rates by 30%, reaching 60% in some departments.
AWS built a solution that goes beyond scheduling appointments to review a patient’s medical history before a doctor’s visit, providing clinicians with a concise overview and insights into active conditions, recent events, trends over time, chronic conditions, and more that may be relevant to both closing care gaps and accurately billing.
AWS has expanded its Ambient Scribe feature to transcribe doctor-patient conversations and create draft clinical notes for providers to review in real-time. The solution also generates patient-friendly post-visit summaries for providers to review and medical codes for clinicians to bill. Each code is linked to source evidence for auditing.
Amazon One Medical clinics have been testing the solution, and ambient documentation now has more than 1 million visits, strong clinician adoption, and is being used regularly on a weekly basis, Aubrey wrote. Amazon One Medical plans to expand into intelligent medical coding this year.
Netsmart, an EHR company serving community-based providers with more than 1,300 customer organizations, also integrated Amazon Connect Health and reported a 275% increase in ambient document adoption.
AWS also claims strict security, patient privacy, and data integrity standards. Aubrey said in a blog post that AWS has 130 HIPAA-eligible services and certifications for global IT and compliance standards.
The company uses something called evidence mapping. This is the ability to connect any output generated by the AI to its precise source, “whether it’s ambient conversation records, patient medical records, or billing guidelines,” Aubrey wrote.
“If an AI-generated summary says, ‘Patient reports poor dietary habits,’ a clinician can click to hear the exact moment of the conversation in which it was discussed. This transparency supports faster and more secure reviews so clinicians can audit, adjust, and make final decisions with confidence,” Aubrey wrote.
Amazon Connect Health leverages specialized supervised fine-tuning and reinforcement learning techniques on healthcare-specific datasets and guidelines. AI capabilities evaluate model performance in multiple stages to ensure safety and accuracy. This includes LLM-as-judge-based assessments, where one AI system checks the work of another, and clinician-driven checks, Aubrey wrote.
