Understand performance, stress, and communication in real clinical environments.
GRW Healthcare combines validated burnout assessment with 468-landmark facial action coding to give clinical leaders a dual-signal view of team wellbeing -- without surveillance, without video storage, and without guesswork.
The Challenge
Healthcare has a visibility problem
Clinician wellbeing directly impacts patient outcomes. But the signals that matter most -- stress, fatigue, communication breakdowns -- remain invisible until it is too late.
of healthcare workers report symptoms of burnout. Traditional surveys catch it months too late.
Miscommunication between care teams is the leading cause of preventable adverse events in hospitals.
of clinical decisions are made under significant cognitive load, increasing the risk of decision fatigue and errors.
Burned-out clinicians are associated with up to 2.4x higher rates of patient safety incidents.
How It Works
From recording to actionable insight in under two minutes
A lightweight, non-invasive workflow designed to fit into existing clinical routines without adding burden.
Upload or Record
Clinicians record a brief 60-90 second session reflection directly in the browser. No app install required. Video is processed entirely client-side and never leaves the device.
AI Analyzes Dual Signals
Our engine maps 468 facial landmarks frame-by-frame using MediaPipe FaceMesh, simultaneously scoring against the Copenhagen Burnout Inventory (CBI) self-assessment to produce a dual-signal analysis.
Dual-Signal Insight
Where self-report and biometric signals align, confidence is high. Where they diverge, the system flags it for deeper exploration -- surfacing patterns the clinician may not consciously recognize.
Improve Outcomes
Clinical leads receive anonymized, aggregated dashboards. Individual clinicians get private, actionable recommendations. Over time, longitudinal trends reveal the real trajectory of team wellbeing.
Features
Purpose-built for clinical environments
Every feature is designed to respect clinical workflows, protect privacy, and surface insights that actually improve outcomes.
Session Analysis
Frame-by-frame facial action unit analysis across 468 landmarks. Generates composure, presence, and cognitive load scores from a single 60-90 second session recording.
Burnout Assessment (CBI)
Validated Copenhagen Burnout Inventory scoring integrated alongside biometric analysis. Produces a research-backed self-report signal that complements facial coding data.
Divergence Detection
Flags cases where self-reported wellbeing and biometric signals diverge. When someone says "I'm fine" but their micro-expressions suggest otherwise, the system highlights the gap for reflection.
Longitudinal Tracking
Track individual and team-level trends over weeks and months. Spot early warning signs of declining wellbeing before they become critical incidents or attrition events.
Scenario Tagging
Tag sessions by context: post-shift debrief, difficult patient interaction, handover fatigue, night shift recovery. Compare signals across scenarios to identify systemic stressors.
Institutional Reporting
Anonymized, aggregated dashboards for clinical leaders. Identify department-level patterns, track intervention effectiveness, and generate compliance-ready reports.
Scientific Foundation
Built on validated research, not marketing claims
Every signal we surface is grounded in peer-reviewed methodology. We report confidence levels for all outputs and use language like “what this likely indicates” rather than making diagnostic claims.
Copenhagen Burnout Inventory (CBI)
Kristensen et al., 2005A validated, peer-reviewed instrument measuring personal, work-related, and client-related burnout across three distinct dimensions. Used in over 500 published studies globally.
Facial Action Coding System (FACS)
Ekman & Friesen, 1978The gold standard taxonomy for classifying visible facial movements into Action Units (AUs). Our system maps 468 landmarks per frame using MediaPipe FaceMesh.
Cognitive Load Theory
Sweller, 1988Framework for understanding how mental effort affects information processing and decision-making under pressure -- directly applicable to high-stakes clinical environments.
Yerkes-Dodson Performance Model
Yerkes & Dodson, 1908The empirical relationship between arousal and performance, demonstrating an inverted-U curve. Moderate stress optimizes performance; too much or too little degrades it.
Our commitment to responsible AI
GRW Healthcare does not diagnose medical conditions, predict clinical performance, or make employment-affecting determinations. All outputs include confidence intervals and are explicitly framed as observational signals intended to support -- not replace -- professional clinical judgment. Our system is designed as a wellbeing support tool, not a surveillance instrument.
Output Preview
Actionable intelligence, not raw data
Every analysis produces a structured report with severity bands, confidence scores, and contextualized recommendations.
Moderate alignment. CBI self-report suggests lower burnout than biometric indicators. This may indicate an early-stage pattern worth monitoring.
Cognitive load indicators are elevated post-shift. Consider reviewing workload distribution or scheduling recovery time before the next high-acuity block.
Signal divergence (24%) suggests a gap between self-assessed wellbeing and observable indicators. A brief check-in conversation may surface underlying stressors.
Composure under clinical pressure remains within a functional range. Communication markers suggest effective handover behaviors despite cognitive load.
Privacy & Security
Privacy is not a feature. It is the architecture.
Healthcare data demands the highest standard of protection. Our system was designed from the ground up to minimize data exposure while maximizing clinical insight.
Client-Side Processing
All video analysis runs directly in the browser using TensorFlow.js and MediaPipe FaceMesh. Video frames are processed locally and never transmitted to our servers.
Zero Video Storage
We never store, upload, or transmit video files. Only derived numerical scores and aggregated metrics are persisted. The raw biometric data is discarded after analysis completes.
Row-Level Security
Database access is enforced at the row level via Supabase RLS policies. Clinicians can only access their own data. Administrators see aggregated, anonymized views only.
PIPEDA-Ready Architecture
Designed to align with Canadian federal privacy legislation (PIPEDA) and provincial health information acts. Consent workflows, data minimization, and right-to-delete are built in.
End-to-End Encryption
All data in transit is encrypted via TLS 1.3. Data at rest is encrypted using AES-256 through Supabase managed infrastructure. API keys and secrets are stored in environment variables, never in client code.
Audit Logging
Every data access event is logged. Administrators can review who accessed what and when, supporting institutional compliance requirements and incident investigations.
Pricing
Assess for free. Pay only for your report.
Take the full 19-question CBI assessment with dual-signal biometric analysis at no cost. When you are ready for your personalized AI-generated report with clinical recommendations, unlock it for a one-time fee.
Assessment
- Full 19-question CBI assessment
- Real-time biometric facial analysis
- Subscale scores across 3 dimensions
- Severity band classification
- Suppression signal detection
- Confidence score
Full Report
- Everything in Free, plus:
- AI-generated clinical narrative
- 3-5 personalized recommendations
- Biometric divergence deep-dive
- Longitudinal trend analysis
- Evidence-based resource library
- Downloadable PDF report
- Email delivery within 60 seconds
Enterprise
- Unlimited assessments & reports
- Team-wide dashboards
- Anonymized aggregate reporting
- Department-level breakdowns
- Custom SSO / SAML integration
- Dedicated account manager
- API access for EHR integration
- PIPEDA & PIPA compliance docs
One burned-out clinician leaving costs $50,000-$100,000 in replacement costs. A $29 report that helps you intervene early pays for itself hundreds of times over.
Get Started
Take your burnout assessment today
Complete the 19-question CBI assessment with real-time biometric analysis. Create an account to access your full AI-generated report with personalized clinical recommendations.
Deploying for your whole team?
Enterprise plans include unlimited assessments and reports, anonymized aggregate reporting, department-level analytics, and a dedicated account manager.
Contact Sales