AI Triage & Symptom Checker

Route patients to the right level of care. Proven across 2.6 million consultations in Rwanda.

AI triage system
Want to improve patient triage in your health system?

Overview

Health systems struggle to direct patients to the right level of care. People with minor ailments overwhelm emergency departments, while those with serious conditions delay seeking help.

Our triage system reaches patients via USSD, SMS, web, or mobile app. It was built and validated across 2.6 million consultations in Rwanda, reducing unnecessary visits while getting urgent cases seen quickly.

You can deploy it as a consumer-facing symptom checker, or plug it into your telemedicine workflows to pre-assess patients before they speak to a clinician.

AI Triage Capabilities

Proven technology serving millions of patients

Symptom Checker

Symptom Checker

A conversational tool that helps patients understand their symptoms and find the right care. Works through websites, apps, and messaging platforms.

Patients describe symptoms in their own words, not medical jargon. The interface offers both free text and guided options, so it works for people of all ages and literacy levels.

The system looks at location, insurance, and nearby facilities to give practical next steps. It might recommend self-care, or point to the nearest in-network provider.

Patients can book an appointment, request a callback, or start a video consultation in one click. The provider receives a structured summary of the assessment.
AI Triage

AI Triage

Assesses patient symptoms and routes them to the right care level. Built on technology proven in Rwanda's national healthcare system.

The algorithms identify high-risk cases that need emergency care, while recognising when symptoms can be managed at home or at a lower-level facility.

Questions change based on each patient's answers, so the conversation feels natural while collecting the clinical data needed for a good assessment.

Runs on smartphones, basic SMS, and USSD. Rural households with limited connectivity can still access quality health guidance.

Try It Yourself

See how patients interact with the triage system via SMS, USSD, or smartphone

9:41TeleClinic App100%
Welcome to Health Triage
Try me!
Clinical Audit Log

Triage Decision Trail • Session ID: Not Started

No triage session active

Start using the phone to begin audit trail

How It Works

Transparent clinical decision-making you can audit and customise

AI Clinical Decision Tree

After each patient response, the AI model updates diagnostic probabilities and determines the next most informative question. Each answer systematically narrows the diagnostic space and builds clinical confidence.

Efficiency & Diagnostic Confidence

The algorithm selects questions that maximize diagnostic value while minimizing patient consultation time. The process typically requires 5-10 targeted questions, with early termination for clinically obvious cases.

Clinical Safety Standards

Patient safety is never compromised for efficiency. The final questions systematically exclude critical warning signs and emergency indicators.

Local Health System Integration

The system is specifically designed for your country's healthcare infrastructure by national clinical experts. They define appropriate care pathways and mandatory safety protocols.

AI Clinical Decision Tree

After each patient response, the AI model updates diagnostic probabilities and determines the next most informative question. Each answer systematically narrows the diagnostic space and builds clinical confidence.

Efficiency & Diagnostic Confidence

The algorithm selects questions that maximize diagnostic value while minimizing patient consultation time. The process typically requires 5-10 targeted questions, with early termination for clinically obvious cases.

Clinical Safety Standards

Patient safety is never compromised for efficiency. The final questions systematically exclude critical warning signs and emergency indicators.

Local Health System Integration

The system is specifically designed for your country's healthcare infrastructure by national clinical experts. They define appropriate care pathways and mandatory safety protocols.

Model Context
{
  "sessionId": "triage-demo-session",
  "patient": {
    "age": 32,
    "gender": "female",
    "insurance": {
      "provider": "Rama",
      "policyNumber": "RW-2024-789456123",
      "packageType": "Mutuelle de Santé - Category 2",
      "type": "public"
    },
    "diagnoses": [
      {
        "icd11Code": "QA02.Y",
        "description": "Essential hypertension - Primary high blood pressure without known cause"
      },
      {
        "icd11Code": "5A11",
        "description": "Type 2 diabetes mellitus - Non-insulin dependent diabetes with metabolic syndrome"
      }
    ]
  },
  "currentStep": 1,
  "currentNodeId": "start",
  "nodeType": "question",
  "isComplete": false,
  "patientJourney": [
    {
      "step": 1,
      "nodeId": "start",
      "type": "question",
      "question": "Welcome to Rwanda Health Line. What is your main health problem today?"
    }
  ],
  "currentStatus": "in_progress",
  "nextAction": "await_user_response"
}
React Flow

How We Ensure Clinical Soundness

Three levels of rigour that make health AI safe and locally appropriate

Clinical Evidence

Clinical Evidence

We build the medical logic on proven foundations. WHO clinical protocols provide the starting point, then we layer on health data specific to your country. Local clinicians validate everything and continue refining the algorithms after launch.

We start with established clinical decision trees from IMCI, IMAI, and ETAT. These provide validated symptom-to-diagnosis mapping, age-specific thresholds, and treatment pathways that have been tested globally.

Health data from your country adjusts for regional disease patterns, facility capacity, and seasonal variation. This makes the baseline relevant to local conditions rather than global averages.

If you have historical consultation data, we analyse it to fine-tune recommendations based on actual outcomes, referral patterns, and what has worked in your specific context.

Local clinicians review the algorithms before launch and continue to refine them through accuracy checks, adverse event monitoring, and protocol updates as guidelines evolve.
Local Adaptation

Local Adaptation

Medical terminology is precise, and precision matters even more in translation. We use a three-stage process: medical professionals translate the content, local clinicians verify clinical accuracy, and real users test it before rollout.

Native speakers with healthcare backgrounds translate all content. They maintain consistent terminology and ensure clinical meaning is preserved across every screen and message.

Local clinicians check symptom descriptions, urgency levels, and cultural fit. They catch nuances that even good translators might miss, like local names for symptoms or conditions.

Before launch, real users work through the system. We test for comprehension, workflow fit, and anything that feels unnatural or confusing in practice.
Sovereign Deployment

Sovereign Deployment

You own the system entirely. It runs on your infrastructure, patient data never leaves your country, and there are no dependencies on external services. If our company disappeared tomorrow, your system would keep running.

Unlike black-box LLMs, the decision trees are fully transparent. Every recommendation traces back to specific clinical logic, with no hallucinations or unexplainable outputs.

The system runs in your data centers. You control the servers, comply with local data protection laws, and own the source code outright.

No external API calls, no foreign cloud services. The system works offline and you manage all updates on your own schedule.

Get Started

The system runs standalone or integrates with your existing telemedicine platform. Here's what implementation looks like:

  • Start with a pilot in one region or facility type
  • We adapt decision trees to your local clinical guidelines
  • Translate content into local languages with clinical validation
  • Integrate with your patient registration or health information systems
  • You own the code and data, and can update protocols independently

Get in touch to discuss how this could work in your context.

Ready to route patients to the right level of care?

Our triage system was validated across 2.6 million consultations in Rwanda. Let's discuss how we can adapt it to your clinical guidelines and local languages.