Amazon Web Services on Thursday unveiled a new artificial intelligence-powered platform aimed at improving patient access to healthcare while reducing administrative workloads for providers.
The system, called Amazon Connect Health, uses agentic AI to integrate with electronic health record systems used by clinicians for tasks such as patient verification, appointment scheduling, compiling medical histories, clinical documentation and medical coding, AWS said in a blog post.
Designed to operate around the clock, the platform can instantly schedule appointments and automatically escalate complex cases to healthcare staff when human intervention is required. It also uses specialised learning techniques trained on healthcare-specific datasets and medical guidelines.
To ensure reliability and patient safety, the platform undergoes multi-step evaluations of model performance, including clinician-in-the-loop reviews to verify accuracy.
According to AWS, UC San Diego Health, which has already deployed the system, reported saving about one minute per call and reducing patient call abandonment rates by up to 60%.
The technology can also transcribe doctor–patient conversations during consultations, draft clinical notes in real time for provider review, and generate simplified summaries for patients.
Amazon Connect Health includes a feature called “evidence mapping,” which links AI-generated outputs to their original sources — such as call transcripts or medical records — to improve transparency and traceability.
Meanwhile, Amazon One Medical has already used the documentation capabilities for more than one million patient visits, with clinicians widely adopting the system and using it regularly each week.
