Microsoft Dragon Copilot
Also known as: DAX Copilot, Nuance DAX, Dragon Ambient eXperience, Dragon Medical One, Microsoft Dragon Copilot
Microsoft's ambient clinical documentation and workflow assistant, merging Nuance DAX with Dragon Medical One, generating specialty specific notes inside Epic and Oracle Health with a required clinician review step.
Dragon Copilot is Microsoft's clinical documentation and workflow assistant, and the largest installed base in ambient clinical documentation. It was formed in March 2025 by merging the Dragon Ambient eXperience line, which Microsoft acquired with Nuance in 2022 for $19.7 billion, with Dragon Medical One voice dictation under a single brand. Microsoft reports more than 600 healthcare organisations deployed on the ambient product, concentrated in large health systems, and the wider Dragon Medical family is used across a large majority of United States hospitals.
The capability set is ambient capture of multi party, multilingual conversations and orders at the point of care or through on demand recordings, converted into specialty specific customisable notes. Its structural advantage over the independent scribes is depth of electronic health record integration: for health systems on Epic the note is generated inside the chart environment with no copy and paste, and Oracle Health is also supported. The clinician always reviews and finalises before anything enters the record, and Microsoft states plainly that clinicians remain responsible for reviewing, correcting and approving medical record content.
Through 2026 the product has been repositioned from a documentation tool toward a clinical workflow platform. HIMSS 2026 brought multilingual support, ICD-10 coding assistance, Microsoft 365 integration and a partner app marketplace where third party developers build specialised applications that plug directly into the Dragon Copilot workflow. Access is role based, with distinct paths for physicians, nurses and radiologists, delivered through web, mobile, desktop and supported EHRs.
The honest picture on outcomes is mixed and worth stating. A peer reviewed study scored generated notes highly on accuracy and completeness, while a separate NEJM AI study at Atrium Health covering 112 clinicians found documentation time reduction of 1.7 percent that was not statistically significant, and concluded that widespread implementation in its then current form was unlikely to produce appreciable productivity gains. Deployment is also an enterprise project: initial technical setup is short, but full rollout including approvals, training and workflow integration is commonly reported at three to six months.
Vendor details
Canonical URL
https://www.microsoft.com/en-us/health-solutions/clinical-workflow/dragon-copilot
Category
Healthcare agent
Company status
acquired
Use cases & customers
Target customers
Deployment options
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 generates the note inside the chart environment, which is its main structural advantage over independent scribes.
Documentation needs differ between physicians, nurses and radiologists. Role based access gives each group its own workflow, structure and deployment path rather than one generic note format.
You are a 25 bed rural hospital with no dedicated IT staff. Microsoft's Rural Health Resiliency Program with Pivot Point Consulting offers 60 percent off list with EHR integration, governance and training bundled into the engagement.
Sources & related URLs
Research sources
Agentic Index coverage score
9.5 / 14 capabilities · 68%
| Integrations & Tool CallingDeep EHR integration with Epic including Haiku and with Oracle Health, generating notes inside the chart environment without copy and paste, plus Microsoft 365 integration and a partner app marketplace. Microsoft Dragon Copilot product page 2026-08-05 | Full |
|---|---|
| Workflow OrchestrationMicrosoft describes AI apps and agents that optimise clinical workflows beyond note capture, covering orders at the point of care, referral letters, after visit summaries and routine task automation, with role based workflow paths. 2026-08-05 | Full |
| Knowledge Grounding & RAGSurfaces critical patient information into the assistant and applies ICD-10 coding assistance, grounding output in encounter and chart context, but no retrieval over a customer knowledge corpus or citation mechanism is documented. 2026-08-05 | Partial |
| Human Oversight & GuardrailsOutput is explicitly a draft for clinician review, nothing reaches the chart unfinalised, and Microsoft states plainly that clinicians remain responsible for reviewing, correcting and approving medical record content. Structural rather than optional. Microsoft Learn 2026-08-05 | Full |
| Security, Identity & GovernanceBuilt on Microsoft's secure cloud with stated trusted compliance and enterprise grade security, sold into an enterprise governance stack, with governance frameworks bundled into deployment engagements. No product specific certification list was retrieved, so graded full at medium confidence on the platform posture. 2026-08-05 | Full |
| Observability & AuditabilityEnterprise governance frameworks accompany deployment and usage is metered per encounter for billing, which implies activity tracking, but no audit trail of AI generated content, reasoning transparency or quality analytics is documented at product level. 2026-08-05 | Partial |
| Memory & State PersistenceFinalised notes persist into the electronic health record and customisable specialty templates persist per user, but no cross session agent memory or accumulating clinician preference model is documented. 2026-08-05 | Partial |
| Deployment & Data ResidencyRuns on Microsoft's cloud with EHR embedded delivery and buyers are advised to confirm regional support during evaluation, but no self hosted option and no documented data residency selection at product level. An independent review reports it is not PIPEDA compliant, which is a competitor sourced claim and unverified. 2026-08-05 | Partial |
| Prebuilt Agents, Templates & PacksSpecialty specific customisable note templates, purpose built role based workflows for physicians, nurses and radiologists, and a partner app marketplace of prebuilt specialised applications. 2026-08-05 | Full |
| Triggers & Channel CoverageAmbient capture of multi party, multilingual conversations at the point of care plus on demand recordings, accessed through web browser, mobile app, desktop app and supported EHRs, with separate access paths per clinical role. 2026-08-05 | Full |
| Model Flexibility & RoutingBuilt on Microsoft's own conversational, ambient and generative AI stack with no customer facing model selection, routing policy or bring your own key capability documented. 2026-08-05 | Unable to verify |
| APIs, SDKs & MCP ExtensibilityA partner app marketplace lets third party developers build specialised applications that plug directly into the Dragon Copilot workflow through Microsoft Marketplace, alongside Microsoft 365 integration. Extensibility by other vendors rather than only by the customer. 2026-08-05 | Full |
| Testing, Debugging & OptimizationThe product has been externally evaluated, with a peer reviewed study scoring note accuracy and completeness at 46.91 of 50 and an NEJM AI study at Atrium Health across 112 clinicians finding a statistically insignificant 1.7 percent documentation time reduction, but no vendor supplied evaluation harness, quality dashboard or regression tooling is documented. 2026-08-05 | Partial |
| Browser & Computer UseNo browser control or computer use capability. The product captures audio and writes into the EHR through integration rather than by operating interfaces. 2026-08-05 | Unable to verify |
The Agentic Index coverage score grades every vendor Full, Partial or Unable to verify 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 by Microsoft; authorised resellers list roughly $369 to $600 per provider per month, one with a $700 setup fee on a 12 month term
per user licence plus per encounter usage
What is public
The licensing structure and the May 2026 changes are documented, as is pay as you go consumption unit pricing. Actual per user and per encounter rates are not.
Billing mechanics
A per user licence combined with usage based consumption. From 1 May 2026 Microsoft simplified the model: the per user licence price was reduced, the separate Physician Practice offer was discontinued with all customers moving to the standard per user licence, all enterprise capabilities were folded into that single licence, ambient encounters became billed per encounter rather than per generated note, and generative AI became billed per encounter rather than per individual AI action.
Cost watchouts
A Dragon Medical One licence is reported as a prerequisite for full feature access, setup fees apply through some resellers, and full enterprise rollout including approvals, training and workflow integration is commonly reported at three to six months
Variable cost rationale
Consumption is billed per encounter on top of the seat licence, so cost scales directly with clinical volume rather than headcount, and different features consume different unit amounts without published rates.
Additional watchouts
Price transparency is the sharpest contrast against the independent scribes in this lane, several of which publish self serve rates an order of magnitude lower. Buyers should also weigh the independent evidence: an NEJM AI study at Atrium Health across 112 clinicians found a statistically insignificant 1.7 percent documentation time reduction, which cuts against the productivity case at enterprise pricing.
Overage / add-ons
Usage based capabilities are metered; pay as you go consumption is documented at $0.01 per consumption unit for eligible ambient and generative AI features, with different features consuming different unit amounts
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 authorised resellers
Commercial notes
The May 2026 restructuring was explicitly a response to feedback that pricing complexity was limiting adoption, which is a rare public admission and useful context for anyone benchmarking scribe pricing. Separately, the Rural Health Resiliency Program with Pivot Point Consulting offers 60 percent off list to eligible independent facilities with EHR integration, governance and training bundled, which is a materially different entry point from the enterprise motion.
Key ambiguities
Microsoft publishes licensing mechanics but not rates, and the two authorised reseller figures differ by more than 60 percent, so the effective per provider cost cannot be pinned down without a quote. How the May 2026 per user price reduction lands against those reseller rates is also unknown.
Missing data
Microsoft's own per user rate, the per encounter rate, and which capabilities consume how many consumption units.
Related vendors
- Abridge — Enterprise ambient AI documentation platform, Best in KLAS 2025 and…
- Adonis — AI orchestration platform for revenue cycle management whose…
- AKASA — Generative AI for the healthcare revenue cycle with institution…
- Alaffia Health — Agentic AI platform for health plan claims operations that pairs…
- AlethianAI — AI native multi agent healthcare platform that automates clinical…
- Almanac Health — Research validated clinical AI platform whose Almanac Copilot acts…
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.
- Sunoh.ai8.5 / 14A lighter documented profile than Microsoft Dragon Copilot
- Hippocratic AI9.0 / 14Fuller documented coverage on Testing, Debugging & Optimization
- mdhub8.0 / 14A lighter documented profile than Microsoft Dragon Copilot
- Prosper AI10.0 / 14Fuller documented coverage on Observability & Auditability and Deployment & Data Residency
- Scribeberry10.0 / 14Adds documented Browser & Computer Use
- Suki8.0 / 14A lighter documented profile than Microsoft Dragon Copilot
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded