NHS - Automated Scribe for Clinician
Fabric Score 3.4
PublicHealthNational Health Service UKGenAI modelUser / Operator / Front-line Staff ×1
Workflow Diagram
Fabric Score
Task
During a patient–doctor consultation, the AI system generates notes and a referral letter (if needed) for the doctor to review and use.
Intent
The AI system improves note-taking efficiency and thoroughness.
AI Workflow
Input
With patient approval, the patient-doctor consultation is recorded and passed to the AI system.
Process
The AI system generates consultation notes and predicts if a referral is needed; if so, the AI system drafts a letter.
Output
Clinician reviews/edits notes and adds them to the chart; if a referral is appropriate, clinician reviews/edits the letter before sending.
Human Oversight Level
Human-Led with AI Assistance
Institutional Oversight Examples
- •Digital Clinical Safety Practice
- •Data Protection (regulation)
- •Equality and health impact assessment
Risk
Clinicians could become over-reliant on the AI system for documentation.
Output Modification Telemetry
- Factual errors / hallucinations70%
- Unfaithful outputs (policy, tone, or style misalignment)10%
- Missing or incomplete information10%
- Irrelevant information5%
- Internal inconsistency5%
Transversal Metrics
Grouped by Fabric dimension.
Efficacy
Accuracy85 %
Efficiency
Modification Rate30 %
Modification Time10 min/output
Verification Time5 min
Rejection Rate10 %
Operational Friction5 hrs/wk
Implementation Overhead300 hrs
Governance Overhead300 hrs
Time to Launch20 wks
Value
Effort Reduction90 %
Risk
Reliability3 incident/month
Autonomy80 %