Amazon Brings Its AI-Powered Healthcare Assistant Directly to Users Through Its Website and App

Amazon's $3.9 billion purchase of primary care company One Medical back in 2023 raised plenty of eyebrows. Now, nearly two years later, the e-commerce giant is showing one of its clearest signals yet of what it actually intends to do with that acquisition — putting an AI-powered health assistant directly inside the hands of anyone who shops on its platform.

Health AI Moves Beyond One Medical's Walled Garden

Until now, Amazon's Health AI assistant was tucked away inside the One Medical app, accessible only to that service's members. That changes with Tuesday's announcement: the assistant is now rolling out across Amazon's main website and mobile app, with no Prime subscription or One Medical membership required to use it.

The tool covers a meaningful range of tasks. Users can ask general health questions, get explanations of their medical records and lab results, manage prescription renewals, and book appointments. For those who grant access, Health AI connects to the Health Information Exchange — the nationwide system for sharing patient data — allowing it to deliver answers personalized to an individual's actual medical history rather than generic advice.

When a conversation moves beyond what an AI can handle, the assistant can escalate to a real clinician. Prime members in the United States get up to five direct-message consultations with One Medical providers at no extra charge, covering more than 30 conditions ranging from cold and flu to UTIs, allergies, and hair loss. Non-Prime users can access the same providers through a pay-per-visit model.

The Privacy Question Amazon Hasn't Fully Answered

Sharing intimate health details with an AI system run by one of the world's largest consumer data companies is not a trivial decision, and researchers have been vocal about the risks. Studies and warnings from institutions including Stanford and Duke have flagged that AI companies frequently use user conversations to train their models — sometimes in ways users don't fully understand.

Amazon's position is that Health AI trains on "abstracted patterns without directly identifying information." The company's example: if many users ask about drug interactions, those patterns can improve the model's responses without exposing anyone's name. All interactions, Amazon says, take place within a HIPAA-compliant environment protected by encryption and strict access controls.

What Amazon hasn't provided, however, is specifics — exactly how conversations are encrypted, who within the company can access them, and under what circumstances. TechCrunch has asked Amazon for those details. Until clearer answers arrive, users are being asked to extend significant trust to a platform whose business has historically been built around knowing as much as possible about its customers.

Why This Move Signals a Broader Shift in AI-Driven Healthcare

Amazon isn't operating in a vacuum here. The healthcare AI space has compressed dramatically in recent months. OpenAI launched ChatGPT Health in January, pitching a version of its chatbot specifically tuned for medical questions — a product it says 230 million users a week were already implicitly asking for. One week after that, Anthropic announced Claude for Healthcare. The race to position AI assistants as the first point of contact between patients and the medical system is now very much underway.

What separates Amazon's play from those of OpenAI and Anthropic is infrastructure. Amazon doesn't just have an AI model — it has One Medical's network of providers, a pharmacy operation, and millions of existing customer relationships. Health AI, at its best, functions not as a standalone chatbot but as a connector between digital triage and actual care. That vertical integration is something neither OpenAI nor Anthropic can currently replicate.

Users interested in trying Health AI can sign up at the Amazon Health page and will receive an email notification when access becomes available to them — a phased rollout that suggests Amazon is being deliberately measured about how fast it scales something this sensitive. Whether the convenience on offer proves compelling enough to overcome legitimate privacy concerns will depend almost entirely on how much transparency Amazon is willing to provide before people are asked to hand over their most personal data.