How It Works
How LearnedCare Sentinel works — a transparent, four-step process designed to augment human expertise, not replace it.
Configure & Upload
Paste text or upload a PDF or Word document directly into the tool. Before scanning, set the context: product type (prescription, OTC, device, supplement), target audience (HCP, patient, public), channel (digital, print, email, journal, leaflet), market, and regulatory framework (ABPI, PAGB, CAP, MHRA, FDA, EFPIA). Organisations can pre-configure defaults so users don't have to set these every time.
Multi-Layer Analysis
The AI runs the content through four analysis layers simultaneously: general compliance checks against the selected regulatory framework; inherent risk detection for claims that cannot be made compliant regardless of rewording; a patient safety layer covering communication ambiguity, risk minimisation, and misunderstanding triggers; and cross-framework regulatory mapping to surface any secondary obligations.
Prioritised Risk Report
Findings are grouped by severity — Alert Red, Risk Amber, and Informational — each linked to the specific regulation or guideline it may breach. Inherent risks are flagged separately so reviewers know they require structural changes, not just edits. Patient safety signals are highlighted in their own category. The full report can be downloaded as a formatted HTML document for records or sign-off.
Rewrite, Decide & Audit
For fixable issues, the tool generates a compliant draft rewrite — copyable or downloadable as a text file. The Compliance Officer reviews all findings, records their decision on each, and the outcome is encrypted and stored in a tamper-evident audit trail. Scan history is available across the organisation. Audit records can be exported as a CSV, and usage is tracked against the organisation's subscription limits.
Configure & Upload
Paste text or upload a PDF or Word document directly into the tool. Before scanning, set the context: product type (prescription, OTC, device, supplement), target audience (HCP, patient, public), channel (digital, print, email, journal, leaflet), market, and regulatory framework (ABPI, PAGB, CAP, MHRA, FDA, EFPIA). Organisations can pre-configure defaults so users don't have to set these every time.
Multi-Layer Analysis
The AI runs the content through four analysis layers simultaneously: general compliance checks against the selected regulatory framework; inherent risk detection for claims that cannot be made compliant regardless of rewording; a patient safety layer covering communication ambiguity, risk minimisation, and misunderstanding triggers; and cross-framework regulatory mapping to surface any secondary obligations.
Prioritised Risk Report
Findings are grouped by severity — Alert Red, Risk Amber, and Informational — each linked to the specific regulation or guideline it may breach. Inherent risks are flagged separately so reviewers know they require structural changes, not just edits. Patient safety signals are highlighted in their own category. The full report can be downloaded as a formatted HTML document for records or sign-off.
Rewrite, Decide & Audit
For fixable issues, the tool generates a compliant draft rewrite — copyable or downloadable as a text file. The Compliance Officer reviews all findings, records their decision on each, and the outcome is encrypted and stored in a tamper-evident audit trail. Scan history is available across the organisation. Audit records can be exported as a CSV, and usage is tracked against the organisation's subscription limits.
Technical Architecture
Our platform is built on secure, modern infrastructure designed for healthcare data sensitivity. We utilise isolated processing environments to ensure that document data is transient and never used for model training.
- UK-hosted infrastructure options
- End-to-end encryption for all transfers
- Role-based access control (RBAC)
Integration Capability
The platform is designed to sit alongside your existing content management workflows. We provide secure API endpoints for automated ingestion or a simple drag-and-drop interface for manual review sessions.
Authorization: Bearer <token>
Content-Type: multipart/form-data