Microsoft Dragon Copilot is positioned by Microsoft as an AI assistant that drafts clinical documentation and supports workflow tasks for qualified healthcare professionals. The product combines conversational and ambient AI with generative AI to produce draft clinical output for clinician review. Microsoft's technical documentation describes standalone web, desktop, and mobile apps, embedded EHR experiences, and developer tools for partners. Availability varies by role, version, and region, and administrators may need to enable specific features before users can access them. [1][4][5]

Documented physician workflows

Microsoft's physician-focused documentation describes ambient capture of multi-party, multilingual conversations and orders at the point of care, or through on-demand recordings. Dragon Copilot can convert those recordings into specialty-specific, customizable draft notes. The product also generates recommendations for discrete orders, conditions, and other clinical data beyond narrative notes. [2][4]

The 5.0 release notes describe an enhanced user interface with a unified microphone experience for ambient and dictation workflows, patient session creation from compact view, and intelligent note section selection. The intelligent selection feature analyzes generated content and chooses note sections such as History of Present Illness, Assessment and Plan, or Physical Exam based on whether content exists, while respecting previously set user preferences. [8]

Patient matching, documented in the 5.2 release notes, applies to Dragon Copilot desktop customers with Epic EHR in the US. The feature indicates when a session and EHR patient do not match and offers options to select the correct session. An optional auto-follow setting can open the matching Dragon Copilot session when a clinician switches patients in the EHR. [6]

Nurse experience and flowsheet documentation

Microsoft's nurse-focused materials describe reducing documentation burden, streamlining flowsheet capture, and automating routine tasks. The use cases white paper lists illustrative nursing uses such as ambient capture of clinician-patient interaction or clinician monologue to populate observations in flowsheet documentation, summarization of interactions and principal points, creation of nurse notes, and summarization of pending activities mentioned in a recording. [2][5]

Flowsheet profile management appeared in the 5.2 release as a public preview for Dragon Copilot for Epic embedded webview in the US. Administrators can assign flowsheet profiles directly to users in Dragon admin center or allow users to select from available profiles. Each profile contains a curated set of flowsheet templates and documentation settings such as lines, drains, and airways, without requiring changes to the underlying EHR profile. [6]

The 5.2 release also expanded lines, drains, and airways support to items up to five years old, replacing the previous 120-day limit. The 6.0 release notes indicate that nurse experience features have moved through public preview and general availability stages, with regional scope documented as US for certain capabilities. [1][6]

Administrative configuration requirements

Several Dragon Copilot features require administrator action before users can access them. For internal sources in the 6.0 release, administrators must upload and configure content in Dragon admin center, and users must then enable internal sources under Settings > Copilot Chat. For the enhanced user interface in 5.0, Microsoft-provisioned customers enable the Switch to new design setting in Dragon admin center, while Nuance-provisioned customers contact their Customer Success Manager. Haiku iOS users also require SMART on FHIR and the Enable Dragon webview for iOS setting. [1][8]

Flowsheet profile management in 5.2 requires customer administrators to create and manage profiles in Dragon admin center. Microsoft's documentation notes that starting with Dragon Copilot 5.2, organizations must create at least one Dragon admin center flowsheet profile; otherwise all test cases in AI customization manager show a Needs review status. Guidance is provided through the customer support manager for setup and alignment with the AI customization manager. [6]

Information assistance and internal sources

Microsoft's marketing materials state that information is brought directly into Dragon Copilot, with responses drawn from transcripts, notes, and trusted third-party medical references with citations. The product description also lists coding suggestions, clinical evidence summaries, referral letters, and after-visit summaries as task automation examples. These are vendor claims about intended functionality rather than independently verified outcomes. [2]

The 6.0 release notes document internal sources as a feature for Microsoft-provisioned customers. Organizations can surface institutional knowledge, policies, protocols, procedures, and operational guidance within Dragon Copilot, allowing clinicians to receive grounded, citation-backed answers from organizational content in Copilot Chat. Administrators upload and configure internal sources in Dragon admin center, and users enable them under Settings. [1]

The 6.0 release also documents multi-intent prompts in Copilot Chat for the UK. Users can combine documentation and information assistance capabilities in a single interaction while maintaining shared clinical context. Microsoft's examples include summarizing a plan and using it to draft a patient instructions letter, or editing an assessment and plan to add a diagnosis and using it to draft a referral letter. [1]

Developer and integration options

The Dragon Copilot Developer Kit provides integration tools for partners who want to embed Dragon Copilot functionality in their own desktop and mobile apps. Microsoft's developer documentation describes audio capture options, data exchange through REST APIs for partners, and Copilot capabilities that provide summarized, contextualized responses from internal and external sources during documentation. [3]

The SDK decision guide distinguishes between embedding the full Dragon Copilot web app, integrating ambient and mobile recording into a partner's own UI, and adding dictation at the cursor. Mobile SDKs for Android and iOS handle recording, data management, and secure communication with the Dragon backend, including offline capabilities. A JavaScript SDK supports desktop experiences requiring ambient recording or dictation in native edit controls. [3]

Dragon Copilot APIs for partners include an ambient session service, configuration service, audio service in preview, and ambient audio streaming service. The APIs are secured with access tokens issued by an identity solution, with Microsoft Entra ID recommended. Identity solutions must be added to an allow list, and access tokens must conform to RFC 9068 with publicly available issuer metadata. [7]

Intended use and operational boundaries

Microsoft's use cases white paper states that Dragon Copilot is intended to assist healthcare professionals with documentation automation and workflow optimization. Use is limited to qualified members of an individual's care team working on the customer's behalf as an employee, independent contractor, or medical staff member. Users must be qualified in clinical practice and trained on Dragon Copilot. [5]

The white paper identifies several considerations for organizations choosing use cases beyond the documented examples. The system is not designed for multi-patient group visits or multidisciplinary team documentation in one session. Only languages explicitly listed in product documentation are supported for speech-to-text and ambient summarization in the current locale, and Dragon Copilot is not a translation tool. Real-time collaborative editing by multiple clinicians in the same note is not supported. [5]

Microsoft's documentation states that output from Dragon Copilot or the Azure OpenAI API does not reflect the opinions of Microsoft, and that accuracy and reliability might vary and are not guaranteed. Microsoft disclaims liability for damages resulting from use of or reliance on information provided by Dragon Copilot or the Azure OpenAI API. Organizations must evaluate legal and regulatory obligations when using AI services. [5]