Microsoft Dragon Copilot
Also known as: DAX Copilot, Nuance DAX, Dragon Ambient eXperience
Microsoft's ambient clinical documentation and workflow assistant, bringing Nuance DAX and Dragon dictation together, that drafts specialty specific notes, orders and summaries inside Epic and other EHRs for clinician review, with an AI apps and agents storefront and a developer kit for partners.
Microsoft Dragon Copilot is Microsoft's clinical documentation and workflow assistant, the successor to Nuance's DAX Copilot and Dragon ambient products. It captures multi-party, multilingual conversations and dictation at the point of care and drafts specialty specific notes, orders, flowsheet documentation, referral letters, after-visit summaries and coding suggestions, with separate experiences for physicians, nurses (Epic Rover) and radiologists (PowerScribe One).
Clinicians can ask Copilot questions answered from the transcript, notes and trusted medical references. Everything it generates is a draft: Microsoft requires the clinician to review and validate content before it becomes part of the medical record. It runs inside Epic and other partner EHRs as well as in its own web, desktop and mobile apps.
Since August 2026 US organizations can deploy AI apps and agents for coding, decision support, risk adjustment, prior authorization and screening from a Dragon Copilot storefront in Microsoft Marketplace, and partners build on a developer kit of REST APIs, SDKs and an extension framework.
Access runs through the customer's Entra ID with roles in the Dragon admin center; Microsoft lists HITRUST CSF, SOC 2 Type 2 and ISO 27001 among its certifications and stores data in the customer's country or region (US, Canada, UK, France and Germany for the EU). Pricing is not published.
Vendor details
Canonical URL
https://www.microsoft.com/en-us/health-solutions/clinical-workflow/dragon-copilot
Category
Healthcare agent
Company status
first party product
Use cases & customers
Target customers
Deployment options
Integrations
Embedded in Epic (Epic Rover for nurses), PowerScribe One for radiology and other partner EHRs; summaries delivered into the EHR; connectors and AI apps and agents managed in the Dragon admin center; developer kit with REST APIs and JavaScript, Android and iOS SDKs.
In practice
Your health system runs Epic and the copy and paste step between a scribe tool and the chart is the friction nobody will tolerate. Dragon Copilot drafts inside the EHR workflow and delivers summaries directly into the record for the clinician to sign.
Documentation needs differ between physicians, nurses and radiologists. Separate experiences give each group its own workflow and deployment path rather than one generic note format.
Your organization is in Canada, the UK or the EU and patient data must stay in country. Dragon Copilot stores customer data in the customer's country or region and inside the EU Data Boundary.
Sources & related URLs
Research sources
Agentic Index coverage score
9.0 / 14 capabilities · 64%
| Integrations & Tool Calling | Full |
|---|---|
|
Dragon Copilot runs inside Epic (including Epic Rover for nurses), PowerScribe One for radiology and other partner EHR integrations, delivers generated summaries directly into the EHR, and admins manage connectors in the Dragon admin center. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/aboutread 2026-09-29 |
|
| Workflow Orchestration | Full |
|
Beyond note drafting, organizations deploy AI apps and agents (coding, decision support, risk adjustment, prior authorization, screening) that run inside the Dragon Copilot point-of-care workflow alongside its own generation, and admins manage them per environment, so multiple agents work in one clinical workflow. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/extensionsread 2026-09-29 |
|
| Knowledge Grounding & RAG | Partial |
|
Copilot answers from the encounter transcript and notes, pulls EHR visit diagnoses into notes, and draws on trusted internal and third-party medical references. No maintained retrieval structure over the organization's own knowledge is documented, and the reference sources are not named. Sourcemicrosoft.com/en-us/health-solutions/clinical-workflow/dragon-copilotread 2026-09-29 |
|
| Human Oversight & Guardrails | Full |
|
All generated content must be reviewed and validated by the clinician before it becomes part of the medical record; output is a draft for the clinician's review, with AI-generated sections marked and the source transcript one click away. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/whitepapers/transparencyread 2026-09-29 |
|
| Security, Identity & Governance | Full |
|
Users sign in through the customer's Entra ID tenant (OpenID Connect) and the Dragon admin center assigns user roles per experience plus Environment Administrator and Analytics Reader roles; Dragon Copilot is HITRUST CSF certified with SOC 2 Type 2, ISO 27001, 27017 and 27018, FedRAMP, C5 and HDS listed. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/whitepapers/securityread 2026-09-29 |
|
| Observability & Auditability | Partial |
|
Each encounter keeps its transcript beside the draft, and admin analytics report usage by site, session type and dropped users. No per-encounter record of the steps and actions the AI took is documented; the audit logs in the security white paper record access to Microsoft's own environment rather than the AI's actions. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/whats-new/5-0read 2026-09-29 |
|
| Memory & State Persistence | Not documented |
|
Templates, custom prompts, shared and personal vocabulary and auto-style are settings and instructions the product applies, and finalized notes are the EHR's record, so no agent memory with a scope and lifetime is documented. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/whats-new/5-0read 2026-09-29 |
|
| Deployment & Data Residency | Full |
|
Customer data is stored by the customer's location: US in the US, Canada in Canada, UK in the UK, France, Ireland and Belgium in France (EU), and Germany, Austria, the Netherlands and Switzerland in Germany (EU), within the EU Data Boundary, and never leaves the geography. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/whitepapers/privacyread 2026-09-29 |
|
| Prebuilt Agents, Templates & Packs | Full |
|
Organizations find, deploy and manage AI apps and agents from the Dragon Copilot storefront in Microsoft Marketplace, covering coding and documentation, decision support, risk adjustment, prior authorization, behavioral health screening and preventive care (US), beside separate physician, nurse and radiologist experiences. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/whats-new/5-0read 2026-09-29 |
|
| Triggers & Channel Coverage | Partial |
|
Work starts when a clinician records an ambient conversation or dictates, from the EHR, the web, desktop or mobile apps, or asks Copilot a question. Each of those surfaces is invoked by the clinician, and no autonomous event, schedule or inbound trigger is documented. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/aboutread 2026-09-29 |
|
| Model Flexibility & Routing | Not documented |
|
Microsoft describes LLM functionality and its own speech and generation models, with Azure OpenAI classifiers named only for content filtering; no model provider list or customer choice of model is documented. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/whitepapers/transparencyread 2026-09-29 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
The Dragon Copilot Developer Kit documents REST APIs for partners, SDKs for JavaScript, Android and iOS (DAXKit), an ambient audio streaming service and Entra ID access token requirements, and an extension framework (adaptive cards, configuration package, Marketplace offer) for building AI apps and agents on the platform. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/sdk/get-startedread 2026-09-29 |
|
| Testing, Debugging & Optimization | Partial |
|
User feedback mechanisms feed continuous improvement. That is stated in one line with nothing on how a change is measured for the customer, and Microsoft's own pre-release accuracy tests, red teaming and A/B gating of new models against baselines test the vendor's product rather than giving the buyer an evaluation surface. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/whitepapers/transparencyread 2026-09-29 |
|
| Browser & Computer Use | Not documented |
|
No browser, desktop or computer control by the agent is documented; Dragon Copilot captures audio and works through EHR integrations and its own apps. Sourcelearn.microsoft.com/en-us/industry/healthcare/dragon-copilot/aboutread 2026-09-29 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Pricing
Not published; contact Microsoft or a partner
What is public
No rates, plans or billing basis are published on Microsoft's product page or in the Learn documentation; AI apps and agents are listed in the Dragon Copilot storefront in Microsoft Marketplace.
Billing mechanics
Not published first party.
Variable cost rationale
No billing basis is published first party, so exposure to encounter volume cannot be checked from Microsoft's pages.
Additional watchouts
Confirm which experiences (physicians, nurses, radiologists) and which countries your license covers, and price any Marketplace AI apps and agents separately.
Sales call required
Yes, required for paid access
Free / trial
No published free tier or self serve trial; access runs through Microsoft licensing channels or authorized resellers
Key ambiguities
Microsoft publishes no rates; the effective cost per provider and per encounter needs a quote from Microsoft or a partner, and Marketplace AI apps and agents are priced separately.
Missing data
Microsoft's per user and per encounter rates, and how Marketplace AI apps and agents are billed.
Related vendors
- Abridge — Enterprise ambient AI documentation platform that turns clinical…
- Adonis — AI orchestration platform for revenue cycle management whose…
- AKASA — Generative AI for the hospital mid revenue cycle: Coding and CDI…
- Alaffia Health — Agentic AI for health plan claims operations: utilization…
- AlethianAI — Connected AI agents for medical practices on one patient record: AI…
- Ambience Healthcare — AI platform for health systems pairing an ambient scribe with…
Alternatives to Microsoft Dragon Copilot
The closest documented capability profiles to Microsoft Dragon Copilot among healthcare agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Nabla7.5 / 14A lighter documented profile than Microsoft Dragon Copilot
- Prosper AI10.0 / 14Adds documented Memory & State Persistence
- SuperDial9.0 / 14Adds documented Browser & Computer Use
- Cohere Health10.5 / 14Fuller documented coverage on Knowledge Grounding & RAG and Observability & Auditability
- Corti9.5 / 14Adds documented Memory & State Persistence and Model Flexibility & Routing
- Scribeberry10.5 / 14Adds documented Memory & State Persistence and Model Flexibility & Routing, among others
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded