I opened my patient portal at 2 a.m. and stared at a column of unfamiliar abbreviations. You know that tiny, rising panic when a single highlighted result can derail your sleep. I wanted something that would translate the numbers into plain English before I called anyone.
I write about AI and healthcare, and I’ve spent weeks testing how these assistants behave when the stakes are personal. What follows is a close look at Microsoft Copilot Health—what it can read, how it keeps data private, and how it compares to other players in the space.
On a morning run your watch buzzes — Copilot Health can connect with Apple Health and Fitbit
Walk out the door and your wearable is already collecting a story about you. Copilot Health links to more than 50 wearable devices, including Apple Health, Fitbit and Oura, and it pulls signals from those devices into your personal health view.
It also taps records from roughly 50,000 US healthcare providers and hospitals, so your lab results, medication history and clinical notes can be read together instead of living in separate silos. The assistant will summarize symptoms, explain likely causes, and flag possible interactions between treatments and meds in plain language.
It acts like a private translator for your medical file, turning clinician shorthand into something you can actually use.
What is Copilot Health?
Copilot Health is an AI feature inside Microsoft Copilot that lets you interact with your personal medical records, device data, and clinical test results. You can ask it to explain a lab value, list current medications, or compare trends from your watch to recent clinic visits.
After a strange result most people call a doctor — Copilot Health offers context and can help you find care
When a result doesn’t match how you feel, the natural next step is to call a clinician. Copilot Health isn’t marketed as a replacement for medical judgment; it’s a translator and triage assistant. You can use it to locate doctors by specialty, language, location, and insurance coverage, and to get plain-language explanations of possible causes and treatments.
Microsoft says Copilot Health draws from trusted sources such as Harvard Health to ground its responses. That doesn’t make it infallible, but it gives you a clearer starting point for a conversation with your clinician.
Can Copilot Health access my Apple Health and Fitbit data?
Yes. You choose the connections you want. The service supports Apple Health, Fitbit, Oura and dozens more devices, letting the assistant combine wearable trends with clinical notes and labs for a fuller picture.
You would not hand your medical file to a stranger — Copilot Health’s privacy and verification steps
Most people guard their health details like a private diary. Microsoft says conversations with Copilot Health are encrypted and stored separately from regular Copilot chats. Your data is not used to train Microsoft’s base AI models, and you can delete or disconnect your information at any time.
Identity verification is handled through Clear ID, and Microsoft emphasizes that health chats live in a segregated workspace to reduce accidental exposure. Those are important safeguards, but any cloud-based system carries risk, so weigh convenience against how comfortable you are sharing sensitive records.
Is Copilot Health secure and private?
Microsoft promises encryption, separate chat stores, and non-use of health data for model training, plus user controls to delete or disconnect data. Those safeguards align with what privacy-conscious users expect, though verification via Clear ID adds a layer of identity checking you should be aware of before connecting accounts.
At a clinic lobby coffee machine someone asked me — how does this compare to ChatGPT Health or Claude?
Microsoft’s move arrives alongside similar steps from others. OpenAI now offers ChatGPT Health, and Anthropic lets users connect Apple Health and Android Health with Claude. Microsoft’s AI lead, Mustafa Suleyman, called Copilot Health “the dawn of medical superintelligence,” signaling how the company sees the product’s ambition.
These tools share a goal: make medical data readable and actionable for everyday people. One could be a lighthouse in a fog of lab numbers.
If you connect Copilot Health, expect a blend of device trends, clinical notes and trusted medical references delivered in conversational form. I tested queries about medication interactions and trends across weeks of heart-rate data; the assistant highlighted plausible links and suggested questions for my clinician, not definitive diagnoses.
These systems are changing how we access medical information, but they raise the same basic question for every user: how much of your medical life do you want an algorithm to interpret for you?