Overview
Suki’s ambient documentation technology automatically generates clinical notes from patient-provider conversations. This allows healthcare providers to focus on patient care while Suki captures, transcribes, and structures the conversation into comprehensive clinical documentation in the background.How It Works
Here is a high-level overview of the process:1
Create Session
You initiate a new ambient session for a patient encounter, providing initial context like patient information and desired note sections. The system returns an
ambient_session_id to track the session.API Endpoint: POST
/api/v1/ambient/session/createCreates a new ambient session and returns an ambient_session_id. You can optionally provide:ambient_session_id(UUID format)encounter_id(UUID format)multilingual(boolean, defaults to false)
2
Capture Audio
You stream the conversation audio to the Suki Platform in real-time. Our Automatic Speech Recognition (ASR) converts the speech to text.
3
Process with AI
Suki’s fine-tuned Language Models (LLMs) analyze the transcript, identify clinically relevant information, and organize it into the appropriate sections of the note.
4
Generate Note
The system generates a structured clinical note. You can then retrieve the generated content or the full conversation transcript.
API Endpoint: GET
/api/v1/ambient/session/{ambient_session_id}/contentAfter ending the session, the system automatically processes the transcript and generates structured clinical content organized by LOINC codes. The note generation happens asynchronously in the background.You can retrieve the generated structured clinical note which includes:- Summary: Clinical content sections organized by LOINC codes (e.g., History of Present Illness, Assessment and Plan)
- Structured Data: Structured clinical data such as diagnoses with codes, descriptions, and types
- Use the
cumulativequery parameter (true/false) to get cumulative or snapshot data - Each summary section includes LOINC code, title, content, and source transcripts used to generate it
5
Retrieve Content
You can retrieve the generated content or the full conversation transcript via API.
API Endpoints:
- Get Session Content: GET
/api/v1/ambient/session/{ambient_session_id}/content- Retrieves generated summary and structured data
- Use
cumulativequery parameter (true/false) to get cumulative or snapshot data - See the Get Session Content API reference for details.
- Get Session Transcript: GET
/api/v1/ambient/session/{ambient_session_id}/transcript- Retrieves the full conversation transcript after session completion
- Includes language identification (
lang_id) and timing information
Key Capabilities
Automatic Speech Recognition (ASR)
Accurately converts medical terminology into written text and distinguishes between different speakers (diarization).
Multilingual Support
Processes conversations in over 80 languages with automatic language detection.
Configurable Note Structure
You can define the structure of the clinical note using predefined templates or standardized LOINC codes.
Background Processing
The system runs passively in the background, automatically handling system interruptions to ensure a seamless user experience.
Security and Compliance
All audio and data are encrypted during transmission and at rest. The Suki for partners is HIPAA-compliant.
Diarization
The system automatically distinguishes between different speakers in the conversation, allowing for accurate transcription and note generation.
Supported Clinical Note Sections
Ambient documentation can generate content for various clinical note sections, including:- History of Present Illness (LOINC: 10164-2)
- Assessment and Plan (LOINC: 51847-2)
- Review of Systems (LOINC: 18776-5)
- Physical Examination (LOINC: 29545-1)
- Chief Complaint (LOINC: 10154-3)
Next Steps
- Get Access: Contact our partnership team to get onboarded as a partner.
- Choose Integration: Select the appropriate integration method (APIs, Web SDK, or Mobile SDK) for your application.
- Test and Deploy: Verify your integration before deploying to your production healthcare application.