What is Nabla?
Nabla is an AI-powered ambient clinical intelligence platform that listens to doctor-patient conversations and automatically generates clinical documentation — SOAP notes, ICD-10 diagnosis codes, procedure codes, and EHR-ready entries. Founded in 2018 in Paris by Alexandre Lebrun (former Facebook AI Research engineer) and Dr. Delphine Groll, Nabla addresses what many clinicians consider the biggest drain on their time and satisfaction: clinical documentation. Studies consistently show that doctors spend 1.5-2.5 hours per day on documentation — time that is unpaid (it happens after clinic hours), unfulfilling (it is clerical work, not medicine), and a leading contributor to physician burnout. Nabla's core innovation is that it works ambiently — the doctor does not dictate, type, or interact with the AI during the patient visit. They simply open the Nabla app on their phone or computer at the start of the visit, have a normal conversation with the patient, and within 30-60 seconds of the visit ending, Nabla generates a complete clinical note. The AI is trained on hundreds of thousands of real clinical conversations and understands medical terminology, clinical reasoning, and the structure of medical documentation across multiple specialties. As of 2026, Nabla is used by thousands of clinicians across the US and Europe and has raised significant venture funding (over $30M). The company's mission is to bring joy back to medicine by eliminating the documentation burden that drives burnout and reduces face-to-face time with patients.
Key Features
Ambient Clinical Intelligence — No Dictation Required
Nabla's defining feature is that it requires zero change in physician behavior during the patient visit. The doctor does not dictate, type, use wake words, or follow structured scripts — they simply have a normal conversation with the patient while Nabla listens (with patient consent). The AI separates the clinical conversation from casual chatter, identifies the relevant medical information, and structures it into a complete clinical note. Specifically, Nabla extracts: chief complaint and history of present illness, review of systems, physical examination findings (both what the doctor says and what they mention during the exam), assessment and differential diagnosis, plan (orders, prescriptions, referrals, follow-up, patient education), relevant negatives (symptoms the patient denies), social and family history updates, and medication reconciliation changes. The AI understands conversational context — it knows that when a doctor says "and how long has this been going on?" they are asking about the chief complaint, and when they say "your blood pressure looks great today" they are reporting a physical exam finding. Nabla supports multiple specialties with specialty-specific note templates: primary care, cardiology, dermatology, orthopedics, neurology, psychiatry, pediatrics, OB/GYN, and more. Notes are generated in standard SOAP format (Subjective, Objective, Assessment, Plan) and are ready for the doctor to review, edit if needed, and sign — typically in under 2 minutes per note compared to 5-15 minutes for manual documentation. The ambient approach is also better for patient experience — patients prefer talking to a doctor who is looking at them, not at a computer screen typing.
Automated ICD-10 Coding & Billing Support
Nabla automatically suggests ICD-10 diagnosis codes based on the clinical conversation and the generated note. The AI analyzes the assessment and plan, identifies the conditions discussed, and recommends the most specific billable ICD-10 codes. For example, if the doctor discusses a patient with type 2 diabetes with diabetic neuropathy, Nabla suggests E11.40 (Type 2 diabetes mellitus with diabetic neuropathy, unspecified) rather than the less specific E11.9 (Type 2 diabetes mellitus without complications). The coding suggestions include: primary diagnosis, secondary diagnoses, and relevant Z-codes (body mass index, tobacco use, long-term medication use, etc.). Nabla also suggests CPT evaluation and management (E/M) codes based on the complexity of the visit — accounting for the 2021/2023 E/M coding changes that base coding on medical decision-making or total time. The AI flags when the documentation supports a higher level of coding than what was selected and identifies documentation gaps that could lead to downcoding on audit. For physicians who previously spent time manually looking up codes or relying on generic EHR code suggestions that often default to unspecified codes, Nabla's specificity represents both time savings and revenue protection — more specific coding supports appropriate reimbursement and reduces audit risk.
EHR Integration
Nabla integrates with major EHR systems to push completed notes directly into the patient's chart. Supported EHRs include: Epic (through Epic's App Orchard marketplace), Cerner/Oracle Health, athenahealth, eClinicalWorks, NextGen, Meditech, and others. The integration works through various mechanisms depending on the EHR: API-based integration (for modern EHRs that support third-party app integration — Nabla pushes the note as a draft that appears in the physician's EHR inbox for review and signing), HL7/FHIR integration (for health systems that support standards-based integration), and clipboard/paste integration (for EHRs without modern APIs — the physician copies the note from Nabla and pastes into the EHR, which takes seconds but is less seamless). For Epic users — the most common EHR among Nabla's US customer base — the integration is deep: Nabla appears as an embedded app within the Epic workflow, and notes are synced bidirectionally (patient demographics and problem list come from Epic into Nabla, clinical notes flow from Nabla into Epic). The integration setup is handled by Nabla's implementation team and typically takes 2-4 weeks depending on the health system's IT processes and the complexity of the EHR environment. For individual physicians or small practices using cloud-based EHRs (athenahealth, eClinicalWorks), setup can be completed in days.
Pros & Cons
Pros
- Ambient approach preserves natural doctor-patient interaction: Unlike dictation-based scribes where the doctor must verbalize findings in structured format, Nabla captures information from natural conversation. This results in better patient experience and more authentic clinical documentation.
- Dramatic time savings: Physicians report saving 1.5-3 hours per day on documentation — time that can be redirected to patient care, professional development, or personal life. This is the primary ROI for individual clinicians.
- Specific ICD-10 coding improves reimbursement: Automated coding suggestions that favor specificity over unspecified defaults can measurably improve appropriate reimbursement and reduce audit risk.
- Multi-specialty support with adaptive templates: Nabla adapts its note structure based on specialty — cardiology notes emphasize different elements than psychiatry notes, and the AI adjusts accordingly.
Cons
- Requires patient consent and comfort with recording: Recording patient visits requires consent, and some patients may be uncomfortable with AI listening to their conversations. Clear communication about how the recording is used, stored, and protected is essential.
- Accuracy depends on conversation quality: If the doctor does not verbalize key findings (assuming they will type them later), the AI cannot capture them. Adopting Nabla requires a shift in workflow — doctors must articulate findings verbally that they previously typed silently.
- EHR integration depth varies: While Epic integration is excellent, other EHR integrations are less seamless — some require copy-paste workflows that reduce the time savings.
FAQ
Is Nabla HIPAA compliant?
Yes. Nabla is HIPAA compliant and will sign a Business Associate Agreement (BAA). Patient conversation data is encrypted in transit and at rest, processed in HIPAA-compliant cloud environments, and is not used to train Nabla's general AI models across customers (each customer's data is isolated). Nabla's data retention policies are configurable — organizations can set how long conversation recordings and notes are retained. The platform also supports GDPR compliance for European customers.
How accurate is Nabla's note generation?
Nabla reports that its clinical notes require minimal editing — typically 1-3 corrections per note, primarily involving adding details the doctor did not verbalize or adjusting nuance in the assessment. The company's internal benchmarks show 95%+ accuracy for capturing explicitly stated clinical information. The AI does not fabricate findings — if the doctor does not mention a physical exam element, the note will state "not examined" rather than inventing a normal finding. Physicians are always expected to review and sign notes before they become part of the medical record.
Nabla Pricing & Plans
| Plan | Cost | Includes |
|---|---|---|
| Individual | $199/month | Unlimited visits, SOAP notes, ICD-10 coding, basic EHR integration, mobile app. For solo practitioners. |
| Team | $179/user/month (5+) | Everything in Individual plus: team analytics, admin dashboard, priority support, advanced EHR integration. |
| Enterprise | Custom pricing | Everything in Team plus: deep Epic/Cerner integration, SSO, custom templates, API, dedicated support, SLA. For health systems and large practices. |
Real-World Deployment — What Adoption Looks Like
Nabla's implementation in clinical settings follows a pattern that has been refined across thousands of deployments. The typical adoption timeline: Week 1 — Setup and training: Nabla's implementation team configures the EHR integration, sets up specialty-specific templates, and trains physicians (typically a 30-60 minute session covering how to use the app, how to consent patients, and how to review and sign notes). Physicians often practice with colleagues before using Nabla with real patients. Weeks 2-3 — Adjustment period: Physicians learn to verbalize findings they previously typed silently. This is the primary behavior change — learning to speak the physical exam findings, assessment, and plan aloud so the AI can capture them. Most physicians report that within 10-20 patient visits, this becomes natural. Note review time during this period is higher (3-5 minutes per note) as physicians learn to trust the AI's output and develop an editing rhythm. Week 4+ — Habitual use: By the fourth week, most physicians report that Nabla is integrated into their workflow. Note review time drops to 1-2 minutes per note. Physicians report that patient interactions improve because they are no longer dividing attention between the patient and the computer screen. The key success factors identified across deployments: physician champions (having at least one enthusiastic early adopter who demonstrates the tool to skeptical colleagues), administrative support (protecting time for training and adjustment — physicians forced to adopt without training time report frustration), and realistic expectations (Nabla is a documentation accelerator, not a replacement for clinical thinking — physicians still need to formulate their assessment and plan; the AI just captures it). Common failure modes: physicians who refuse to verbalize findings and expect the AI to read their mind, EHR integration problems that delay note syncing, and lack of patient consent processes (leading to awkward situations where patients are surprised by recording).
Nabla vs Other AI Scribes — How to Choose
| Feature | Nabla | Suki AI | Dragon Medical One |
|---|---|---|---|
| Interaction Model | Ambient — passive listening | Interactive — voice commands | Dictation — speak structured content |
| Behavior Change Required | Minimal — verbalize findings naturally | Moderate — learn voice commands | High — dictate in structured format |
| Patient Experience | Natural conversation, no device visible | Physician speaks to device | Physician dictates after or during visit |
| EHR Capabilities | Note generation + ICD-10 coding | Notes + EHR query + order entry | Direct EHR dictation (any field) |
| Time Savings | ~2 hours/day (notes only) | ~2 hours/day (notes + EHR tasks) | ~1.5 hours/day (typing replaced) |
| Best For | Physicians who prioritize patient interaction and want zero device interaction | Physicians who want voice control over EHR + notes | Physicians who prefer dictation-based workflow with maximum EHR flexibility |
The choice between these tools depends on physician preference and workflow. Nabla is best for physicians who prioritize natural patient interaction and want to minimize technology during visits. Suki is best for physicians who want a voice-driven EHR interface that goes beyond documentation. Dragon Medical One is the most established option with the deepest EHR integration but requires the most structured dictation. Many health systems deploy multiple options, allowing physicians to choose the tool that fits their style — recognizing that documentation preferences are highly individual and no single scribe solution works for every clinician.
Security and Privacy
Nabla is HIPAA compliant and SOC 2 Type II certified. Patient conversation audio is encrypted in transit and at rest. The audio is processed to generate the clinical note and then deleted according to the organization's retention policy — Nabla does not retain audio recordings indefinitely. The AI models are not trained on individual customer data across customers — each health system's data is isolated. For European customers, Nabla is GDPR compliant with data processing in EU-based cloud infrastructure. Nabla will sign a Business Associate Agreement (BAA) for HIPAA-covered entities.
Related Tools
Suki AI
AI voice assistant for clinicians — dictation-based clinical documentation with EHR commands. Alternative approach for physicians who prefer voice-driven interaction.
Dragon Medical One
Cloud-based clinical speech recognition — dictate directly into EHR. The market leader for voice-driven clinical documentation.