AWS has revealed Amazon Connect Health, a healthcare-specific AI service designed to automate patient engagement and point-of-care workflows.
In its product announcement, AWS said the service delivers five agents across patient verification, appointment management, patient insights, ambient documentation and medical coding, with the aim of reducing administrative work for providers and staff.
What the five agents do and where they connect
Amazon Connect Health is built on Amazon Connect, and AWS describes it as a purpose-built healthcare service for patient engagement and point-of-care workflows.
AWS says Amazon Connect Health supports Epic, one of the largest electronic health record platforms in the US, through a dedicated app in the Epic App Market, while other EHR systems, including Oracle Health’s Cerner platform and MEDITECH, can connect through the Amazon Connect Health EHR proxy service and integration partners.
AWS also says the service is currently available in the US East (N. Virginia) and US West (Oregon) regions.
The workflow: before, during and after the encounter
In its About Amazon post, the company said the system can review a patient’s history before a visit, surface active and chronic conditions tied to care gaps and billing, transcribe the clinician-patient conversation during the visit and then generate after-visit summaries and billing codes after the encounter.
AWS claims every insight, note and suggested medical code is linked to source evidence, including conversation transcripts, patient records and billing guidance.
Responsibility, guardrails and the not-a-medical-device line
In the FAQ, AWS states that every insight, note, and medical code is linked to the original source data, while its user guide states that Amazon Connect Health is not a medical device and that healthcare providers retain full responsibility for the accuracy of clinical documentation and patient care coordination.
The same materials say the service includes escalation paths to human staff, automated guardrails and layered evaluation that includes manual clinical review and output validation.
Early user figures AWS is citing at launch
AWS is also trying to support the launch with operating data from early users, though those figures come from company materials and partner statements. AWS says UC San Diego Health, which it says handles 3.2 million patient interactions annually, has saved one minute per call, shifted 630 hours a week from patient verification to direct assistance and reduced call abandonment rates by 30%, rising to 60% in some departments.
It also says ambient documentation now spans more than 1 million visits at One Medical and that Netsmart has recorded a 275% increase in ambient documentation adoption across a network of more than 1,300 client organizations.
Those claims land in a market where administrative workload remains a documented pressure point in clinical practice. The American Medical Association says administrative burden interrupts patient care and contributes to burnout, while peer-reviewed research has found primary care physicians spending more than half of their workday in the EHR during and after clinic hours.
Pricing, availability and what remains in preview
The launch also comes with more pricing detail than AWS initially emphasized in its broader announcement. AWS says ambient documentation is priced at $99 per user per month for up to 600 encounters, with overage charges beyond that threshold, while patient verification is priced at $0.15 per action.
At the same time, AWS says appointment management and patient insights remain in preview, medical coding is in gated preview and patient insights pricing has not yet been announced. Medical coding currently supports ICD-10, CPT and E/M coding, according to the FAQ.