Microsoft launches Copilot Health, an AI service that unifies medical data and analyzes health

Microsoft has launched Copilot Health, a separate protected space inside Copilot where AI gathers a user's medical data into a single profile and helps explain what it means. The waitlist opened on March 12, 2026, and the first rollout is phased, starting with adult users in the US in English.

How the service works

Copilot Health appears as a separate tab in the web version and mobile app of Copilot. Users can create a basic profile, enter their age and sex, and then connect external data sources. The system then analyzes lab results, wearable metrics and medical record data to assemble what Microsoft calls a "holistic story" of a person's health. The idea is that people see not scattered numbers but the links between sleep, activity, symptoms, medications and doctor visits.

Microsoft pitches the product not as a replacement for a doctor but as a tool for preparing for a conversation with one. The service is meant to help people make sense of confusing tests, spot patterns in everyday metrics and formulate questions before an appointment. According to the company, its consumer products Copilot and Bing already answer more than 50 million health questions a day, and Copilot Health is an attempt to turn that demand from general search into a personalized experience.

What data can be connected

The personal layer of Copilot Health is built around connectors that pull medical and everyday information from several sources. At launch, Microsoft names three main data channels and supplements them with external reference material, so answers do not rely on the user's profile alone. The result is a service that aims to be more than a window for uploading files: a place where AI ties dry figures to clinical and behavioral context.

• activity, sleep and vital-sign metrics from more than 50 devices, including Apple Health, Oura and Fitbit;
• electronic medical records and discharge summaries from more than 50,000 US hospitals and medical organizations via HealthEx;
• lab test results via the Function service;
• reference information from vetted medical organizations in 50 countries, with citations and links inside answers.

On top of this, Copilot Health can search for doctors in real time by specialty, location, language and insurance, and it shows answer cards prepared by experts at Harvard Health. That makes the service less a health chatbot than an attempt to combine personal data and external vetted knowledge in one interface. For Microsoft it is also a step toward what the company calls medical superintelligence: a system that could eventually combine the breadth of a family doctor with the depth of a specialist.

Privacy questions

The most sensitive part of the launch is not the model but the rules for handling data. Microsoft stresses that Copilot Health stores conversations and medical information separately from regular Copilot, uses encryption for data at rest and in transit, and lets users disconnect any connected source at any time. The company separately states that data from Copilot Health is not used to train models, which is one of the key trust signals in the consumer health AI market.

"Copilot Health does not replace a doctor.

It helps make time with one more useful."

During a press briefing, Microsoft clarified an important regulatory detail: at launch, the service does not fall under HIPAA as a traditional healthcare provider, because it works directly with consumers, who upload and connect their own data. The company says new features will be released only after clinical evaluation and with explicit labeling, and that the product has already received ISO/IEC 42001 certification for AI management systems. Still, trust will remain the central question: the deeper AI goes into health, the higher the bar, not only for answer quality but also for the service's legal model.

What this means

Microsoft is making the most direct bet yet on consumer health AI, trying to take the space between searching symptoms online and a full consultation with a doctor. If Copilot Health can actually turn medical records, wearable data and lab results into clear conclusions without causing needless anxiety, it could quickly become one of the most practical use cases for personal AI assistants.