Tech
Amazon makes a new bet on healthcare AI, rivaling Microsoft in the doctor’s office
Amazon Web Services is expanding into AI for healthcare, launching a new agentic system that can handle patient calls, document clinical visits and automatically generate billing codes.
Amazon Connect Health, announced Thursday morning, is the first industry-specific extension of the cloud giant’s Amazon Connect system for call centers, which crossed the milestone of a $1 billion annual revenue run rate last year.
It will compete in part with rival Microsoft, which acquired Nuance for $19.7 billion in 2022 and has embedded its DAX Copilot ambient documentation tool into major electronic health record systems. AI scribe startups have raised hundreds of millions of dollars to automate clinical documentation.
Amazon is pitching Amazon Connect Health as a broader solution that spans the full healthcare workflow, from the initial phone call through the post-visit billing code.
The idea is to “not provide just point solutions, point tools, or a collection of capabilities, but think end-to-end about what is the customer problem, and how can we solve it,” said Rajiv Chopra, AWS vice president of Health AI and Life Sciences, in an interview this week.
Early users of the technology include UC San Diego Health, which handles 3.2 million patient interactions annually; One Medical, the Amazon-owned primary care practice that has used the ambient documentation capabilities across a million visits; and Netsmart, which provides EHR software to more than 1,300 community-based healthcare organizations.
Amazon’s move could double as a litmus test for AI adoption in healthcare, where institutions have traditionally been slow to adopt new technology.
A randomized trial published in the New England Journal of Medicine AI in December found that ambient AI documentation (the AI startup Abridge) reduced clinician burnout and cut documentation time by 30 minutes per day per provider. However, hospitals continue to cite concerns about data privacy, the difficulty of integrating AI tools into existing workflows, and unclear return on investment.
Amazon’s new product has five core capabilities: automated patient verification, intelligent appointment scheduling, pre-visit summaries for clinicians, ambient documentation that transcribes and drafts clinical notes during the visit, and automated medical coding for billing.
It integrates natively with Epic, the largest U.S. electronic health records system, and connects to other EHRs through data integration partners. It also connects to AWS HealthLake, Amazon’s cloud-based health data repository, which is getting new agentic capabilities to convert records into standard formats.
Amazon Connect Health comes from AWS’s Applied AI Solutions group, led by Senior Vice President Colleen Aubrey, which is focused on building finished applications for specific industries rather than selling raw cloud infrastructure and tools to developers.
Aubrey, who previously built Amazon’s advertising business, said at AWS re:Invent in December that her team is putting “agentic AI at the heart of everything we do,” describing the technology as “AI teammates” that can work autonomously on behalf of businesses.
Healthcare is the first vertical to get a purpose-built Connect product, but Aubrey signaled that more are in the works. Separately, the group oversees Amazon’s Just Walk Out retail technology and is developing agentic AI tools for supply chain planning and life sciences.
Amazon Connect Health is available in preview starting Thursday.