Seattle, March 6, 2026, 4:32 AM UTC-08:00
Amazon.com’s cloud unit launched a healthcare-focused AI service on Thursday, moving deeper into a market where hospitals are trying to cut phone, paperwork and billing work. Amazon Connect Health uses what AWS calls agentic AI — software that can complete tasks with limited human prompting — inside electronic health records, or EHRs, the digital files clinicians use, to verify patients, schedule visits, write notes and generate billing codes. 1
The timing matters because investors are still asking when Amazon’s huge AI build-out will start producing clearer returns. Reuters reported last month that Amazon plans about $200 billion of capital spending in 2026, and D.A. Davidson analyst Gil Luria said the company “has to invest at these levels just to stay in the race.” 2
Competition in healthcare is moving quickly too. CVS Health and Google Cloud said on Thursday they would launch an AI-powered health platform, while Microsoft built its healthcare administration push around Nuance, which Reuters reported in 2021 was bought to expand cloud tools for providers and clinical documentation. 3
AWS said patient verification and ambient documentation — software that listens during visits and drafts notes — are generally available now, while appointment management, patient insights and medical coding remain in preview. Naji Shafi, AWS’s director of healthcare AI, told MobiHealthNews the system integrates directly with Epic in real time for patient verification and appointment management. 4
AWS is pitching the product as a broader workflow tool, not a single feature. Rajiv Chopra, AWS vice president of Health AI and Life Sciences, told GeekWire the goal was not “point solutions,” and an AWS blog said staff at large health systems can spend up to 80% of call-handling time compiling patient data across disconnected systems. 5
Amazon pointed to early customer results to back the case. It said UC San Diego Health, which handles 3.2 million patient interactions a year, saved a minute per call, shifted 630 staff hours a week from verification to direct patient help and cut call abandonment by 30% on average, reaching 60% in some departments; Amazon also said One Medical has used the documentation feature in more than a million visits. 6
AWS said the service can work 24/7, hand tougher cases to human staff and trace each AI-generated note or code back to the source through what it calls evidence mapping. In practice, that means a clinician can pull up the exact transcript or record behind the output instead of taking the software on trust. 7
But healthcare AI is still a risky sell. Reuters reported in January that ambient scribe tools are already raising privacy, consent and malpractice questions as providers rush them into clinics and try to fit them into existing state and federal rules. 8
The accuracy issue has not gone away either. A February Reuters report on a study in The Lancet Digital Health said AI systems were more likely to repeat medical mistakes when those errors appeared in authoritative-sounding documents; “current AI systems can treat confident medical language as true by default,” Mount Sinai’s Dr. Eyal Klang said. 9
That uncertainty sits beside Amazon’s broader spending test. AJ Bell investment director Russ Mould said the selloff after Amazon’s February spending plan reflected a move away from stocks “where positive surprises may be hard to achieve,” making launches like Connect Health more closely watched than a routine AWS product release. 10
