Amazon, Perplexity and OpenAI Compete to Own the First Click in Healthcare

The same week, lab testing company Function announced a connector that lets its members pipe lab results and clinician-reviewed summaries directly into Perplexity Health.

Those moves follow OpenAI’s ChatGPT Health and Microsoft’s Copilot Health on March 12, bringing the number of major AI platforms with dedicated consumer health products to five in under three months.

In the U.S., 3 in 5 adults used AI tools for health purposes in the past three months. About 70% of AI health conversations occur outside clinic hours, and roughly 1.6 million to 1.9 million messages per week on ChatGPT focus on health insurance questions, according to PYMNTS coverage. The platforms responding to that demand are not building the same product. They are building different layers of the same stack.

Vertical Stack: Amazon Owns the Full Care Workflow

Amazon builds vertically. Its Health AI agent does not hand users off to another system. It handles the question, the appointment, the prescription and the specialist referral inside one product, connected to one provider network.

Health AI answers health questions, explains lab results and medical records, manages prescription renewals and connects users to One Medical providers through message, video or face-to-face visits. The system runs on Amazon Bedrock as a multi-agent architecture: A core agent handles patient communication, sub-agents oversee specific tasks, auditor agents review conversations in real time, and sentinel agents escalate to human providers when needed.

Prescriptions are routed to Amazon Pharmacy or any pharmacy of the user’s choice. One Medical maintains clinical partnerships with Rush University System for Health and Cleveland Clinic for specialist referrals, with Health AI routing patients across those networks.

Horizontal Stack: Perplexity Aggregates Any Source

Perplexity builds horizontally. Perplexity Health does not tie users to a specific provider network. It pulls data from across fragmented health data and surfaces answers from it all at once.

Perplexity Health launches with connectors for Apple Health, electronic health records from more than 1.7 million care providers, and wearable platforms including Fitbit, Ultrahuman and Withings.

A question about resting heart rate draws on that user’s cardiac history, recent bloodwork, and activity logs. Answers draw from clinical guidelines and peer-reviewed journals, with each response connected directly to source material.

Health data is encrypted in transit and at rest, is not used to train AI models and is not sold to third parties. The product is available to Pro and Max subscribers in the U.S. on iOS and at perplexity.ai/health.

Function is a company that provides lab testing and clinician-reviewed biomarker insights. It announced a partnership with Perplexity Health to integrate this data into the AI system, with CEO Jonathan Swerdlin noting, “AI is most powerful when it’s personal,” as the collaboration helps turn fragmented health data into useful insights.

OpenAI and the Scale of the Shift

OpenAI established the category in January with ChatGPT Health as reported by PYMNTS. The usage numbers behind the launch explain the competitive urgency. More than 230 million people globally ask health and wellness questions on ChatGPT every week, according to TechCrunch.

Among U.S. adults who used AI for health in the past three months, 55% used it to check symptoms, 48% to understand medical terms and 44% to learn about treatment options, according to OpenAI.

As PYMNTS covered, AI agents in healthcare are moving from pilots to live deployments across patient engagement, care coordination, clinical documentation and chronic care management.

Amazon operates a vertically integrated model, combining its AI, provider network, pharmacy and billing into a single system. Perplexity takes a horizontal approach, offering one interface that sits across providers, wearables and labs. While OpenAI brings the largest distribution channel, with 230 million weekly health users and a multiyear lead in clinical model development. Each approach is distinct, but together they point to a consumer healthcare system where AI becomes the first point of interaction before any clinician is involved.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.