On March 5, Amazon Web Services launched Amazon Connect Health, an agentic AI solution that simplifies administrative tasks such as scheduling and clinical documentation while speeding up access to care for patients.
Connect Health integrates with EHRs to transcribe doctor-patient conversations, draft clinical notes and generate after-visit summaries.
The burden of a clinician’s administrative workflow — including verifying patient identities and pulling together patient records across multiple systems — can interfere with delivering actual care, explained Colleen Aubrey, senior vice president of applied AI solutions at AWS.
“For health care providers: the work of delivering care shouldn’t be buried under the work of managing it,” Aubrey wrote in a blog post. “Less time spent on preparation, documentation, and billing means more face-to-face time with patients, and health systems that run more efficiently. That’s the gap we set out to close with Amazon Connect Health.”
In addition to connecting to EHRs, Connect Health integrates the functionality of Amazon Connect, which launched in March 2017 as a cloud-based contact center.
“While Amazon Connect is an AI-powered solution for delivering personalized customer experiences across industries, Amazon Connect Health is a new healthcare solution that extends beyond patient experiences to improve workflows for patients, clinicians, and administrative staff,” explained Naji Shafi, health AI director at AWS, in an email interview.
AWS previewed the patient verification and appointment management capabilities of Connect Health at AWS re:Invent 2025 in December.
UC San Diego Health is using Amazon Connect Health to improve patient interactions and reduce call abandonment rates by 30%, and even 60% in some departments, according to Shafi. One Medical is also using Amazon Connect Health features such as ambient documentation, and has now added medical coding.
Connecting agentic AI to EHRs
Connect Health integrates features like appointment management with Epic and more than 100 other EHRs through AWS’ data integration partners. Scheduling and appointment reminders are among the top uses for agentic AI in healthcare.
Along with these integrations, a unified software development kit (SDK) speeds up deployment of Connect Health’s patient engagement features to days instead of months, according to Shafi, adding that the SDK integrates directly into EHR screens and digital front doors without requiring screen and workflow changes.
The patient verification feature, in general availability, allows providers to avoid a multistep process of looking up patients. Instead, they get real-time EHR access.
“It confirms who the patient is, checks their insurance, reviews patient and provider availability — and books the appointment while the patient is still on the line,” Shafi explained. “This also reduces administrative burden by automating scheduling and verification, freeing staff to focus on patient care.”
Building patient insights from EHRs
Amazon Connect Health can draw on internal EHR data as well as external sources such as health information exchanges (HIEs) and health networks to synthesize longitudinal medical history. Rather than producing a generic pre-visit summary, the solution generates personalized insights, aiming to save clinicians time when conducting manual chart reviews and allow them to manage chronic conditions more comprehensively.
“Instead of spending time reviewing charts before a visit, doctors can go straight to the exam room, and spend time actually interacting with patients,” Shafi said. “Pre-visit insights are surfaced automatically.”
AWS also addressed data fragmentation in healthcare by offering connectivity into AWS HealthLake with a “petabyte-scale, FHIR-first persistence layer,” Shafi said.
“With the combination of Amazon Connect Health and AWS HealthLake, customers simply point agents like patient insights to these unified data sources to receive context-driven intelligence grounded in comprehensive patient histories, enabling clinicians to make informed decisions without piecing together information from multiple systems,” Shafi said.
Easing the documentation workflow
Ambient documentation features in Amazon Connect Health produce notes from patient-clinician conversations and automatically format them into existing EHR templates. At the same time, it auto-populates after-visit summaries in an easy-to-read format for patients. It then generates medical codes such as ICD-10, CPT and E&M to make the visits billing-ready within minutes, according to Shafi.
In addition, Amazon Connect Health includes evidence mapping, which links AI-generated output back to an original source, such as an ambient conversation transcript or a patient’s medical records. Physicians can click in a transcript to hear back certain moments from a patient conversation, which helps with auditing and transparency, Aubrey wrote.
The capabilities of agentic AI in platforms such as Connect Health allow AI to act as a teammate rather than simply a tool, Shafi suggested.
“We’re offering multiple capabilities that help reduce administrative burdens, and we are building towards a solution where context flows across [the] entire care journey,” Shafi said.
Brian T. Horowitz started covering health IT news in 2010 and the tech beat overall in 1996.
