TRIBOT - Multilingual AI for emergency care

Transforming emergency care
with AI-powered multilingual triage

Contact us
Trial TRIBOT
Personalise
Busy reception desk with many patients waiting to attend consultation, trying to write medical report in hospital lobby. Diverse people having appointments in waiting room area.

Every second counts in the Emergency Department. Yet, for millions of linguistically diverse patients, language barriers delay care, increase clinical risk, and worsen outcomes.

TRIBOT is a conversational AI triage agent purpose-built to bridge that gap. By combining large language models, clinical reasoning frameworks, and linguistic validation, TRIBOT delivers real-time multilingual interpretation and clinically validated triage classification — supporting safe, timely, and equitable emergency care.

Our vision

A world where language is never a barrier to urgent, life-saving healthcare.

Project at a glance

NHMRC IDEAS Grant

~$980K competitive funding (CIA: Dr Padmanesan Narasimhan), supporting development, validation, and pilot deployment.

Bilingual MVP

English and Arabic prototype validated against the Australasian Triage Scale, with expansion roadmap to additional CALD-priority languages.

Global Partnerships

TRIBOT-GLOBAL bilateral extension under development with Swiss partners (EPFL, ETH Zurich, CHUV, University of Bern).
Try the TRIBOT prototype

Technology

TRIBOT technology overview

TRIBOT is not a translator. It is a context-aware clinical assistant engineered for the most high-stakes environment in healthcare — the Emergency Department.

Watch the TRIBOT Technology Overview

A short video walkthrough of TRIBOT's architecture, workflow, and capabilities.

Core components

Conversational AI & NLP

Natural, human-like multilingual interactions tuned for clinical disclosure, symptom elicitation, and red-flag detection.

TRIBOT-LM

Domain-adapted large language model fine-tuned on bilingual (English / Arabic) emergency triage dialogues.

ACTRI-Bench

Australasian Clinical Triage benchmark for evaluating LLM performance against the Australasian Triage Scale (ATS).

CASMF-Triage

Clinical AI Safety Monitoring Framework purpose-built for conversational triage agents.

TriageSim

Sequential triage simulation benchmark probing anchoring bias and evidence insensitivity in LLM decision-making.

Continuous Learning

Closed-loop feedback from clinician evaluators drives ongoing refinement and safety improvements.

Engineered for Safety

  • Linguistic & Clinical Validation — synthetic triage dialogues co-evaluated by ED clinicians, NAATI-certified interpreters, and applied linguists.
  • Hallucination Mitigation — constrained generation, retrieval-grounded reasoning, and uncertainty-aware deferral to clinicians.
  • Inference-Latency Optimisation — engineered for real-time clinical use, with sub-second response targets in pilot deployment scenarios.
  • Ethical by Design — developed within NSW Health's AI governance framework, with privacy, transparency, and cultural safety embedded from the ground up.
  • Validated Against Gold Standards — aligned with the Australasian Triage Scale and benchmarked against clinician decisions.

Industry & Research Partnerships

CSIRO iPhD scholarship secured with Cogninet Australia as industry partner, focused on AI safety, hallucination mitigation, and clinical safety monitoring. Bilateral SNSF–NHMRC grant under development for the TRIBOT-GLOBAL extension.
Return to top

Applications

Global Scalability, Multi-Sector Impact

TRIBOT is engineered for global scalability and deployment across a range of care settings where linguistic diversity intersects with clinical risk.

Emergency Departments

Reduce triage misclassification, waiting times, and communication-related clinical errors for CALD patients.

Hospitals & Telehealth

Provide 24/7 multilingual interpretation across acute, sub-acute, and virtual care settings.

Aged Care & Community Health

Support linguistically diverse populations with safe, immediate communication tools at the point of care.

Education & Training

Platform for studying AI adoption, clinician–AI interaction, and decision support in complex clinical environments.

Health System Research

Open-science benchmarks (ACTRI-Bench, TriageSim) enabling reproducible evaluation of clinical conversational AI.

Commercial Scale-Up

Expandable across languages and jurisdictions, supporting partnerships with health systems worldwide.

TRIBOT-GLOBAL — International Expansion

Through the TRIBOT-GLOBAL bilateral extension, the program is partnering with leading Swiss institutions — EPFL, ETH Zurich, CHUV (Lausanne University Hospital), and the University of Bern — to adapt and validate TRIBOT for European multilingual emergency care contexts, with co-funding under preparation through the SNSF–NHMRC bilateral mechanism.

 

Looking Ahead

TRIBOT's roadmap extends into wearables, home health monitoring, and multilingual patient education, positioning the program as a foundation for next-generation digital health infrastructure across high-stakes clinical settings.
Return to top

Team

Together, our team brings world-class expertise across clinical medicine, AI engineering, linguistics, and health systems — ensuring TRIBOT is both scientifically robust and socially impactful.
  • Dr Padmanesan Narasimhan (Padma)

    CIA-A · Founder & Lead Investigator
    Senior Lecturer & Lead, Digital Health & AI — UNSW Medicine & Health, School of Population Health; ED Clinician, SWSLHD. WHO-recognised expert in AI and Digital Health with a strong track record in epidemiology, health systems, and emergency medicine.
  • Senior Research Fellow and Post Graduate Coordinator SWS Clinical Campuses, School of Clinical Medicine Roberto Forero
    Roberto Forero
    Senior Research Fellow and Post Graduate Coordinator SWS Clinical Campuses, School of Clinical Medicine
    View Profile
    Aditya Joshi
    Aditya Joshi
    View Profile
    Professor of Interpreting and Translation Sandra Hale
    Sandra Hale
    Professor of Interpreting and Translation
    View Profile
    Jitendra Jonnagaddala
    Jitendra Jonnagaddala
    View Profile
    Adjunct Associate Professor Usman Iqbal
    Usman Iqbal
    Adjunct Associate Professor
    View Profile
    Blank avatar headshot
    Mohammed Mohsin
    View Profile
    Director of Lifelong Learning Anthony Sunjaya
    Anthony Sunjaya
    Director of Lifelong Learning
    View Profile
    Blank avatar headshot
    Assoc Professor Sally McCarthy
    Chief Investigator
    Blank avatar headshot
    Dr Matthew Smith
    Chief Investigator
    View Profile

  • Ms Fang (Sam) Shen
    Associate Investigator
    SWSLHD

    Assoc Prof Andrew Coggins
    Associate Investigator
    Westmead Hospital

    Wayne Varndell
    Associate Investigator
    Prince of Wales Hospital

    Dr Benjamin Harris-Roxas
    Associate Investigator
    UNSW Sydney

    Assoc Prof Holly Seale
    Associate Investigator
    UNSW Sydney

    Dr Chi Ho Chan
    Associate Investigator
    UNSW Sydney

    Mrs Brynn Quick
    Associate Investigator
    Macquarie University

    Dr Patrick San Gabriel
    Associate Investigator
    eHealth NSW

  • Narcis Pasic
    Project Manager (TRIBOT)
    UNSW Sydney

    Michael Earey
    Software Developer (TRIBOT)
    UNSW Sydney

    Joseph Zhu
    Research Assistant
    Data Science & Engineering

     

    Rakibul Hassan Rejon
    Frontend & UX 

    Umayr Mallah
    Recruitment/Project Officer
    UNSW Sydney

    Linju Joseph
    Qualitative Research Expert
    UNSW Sydney

    Suzan Makhloof
    Linguistic Specialist, Arabic Language

    Sam Hoballah
    Linguistic Specialist, Arabic Language

  • Paul (Quoc Dung) Nguyen
    PhD Candidate
    Digital Health & AI — CALD patient outcomes in ED triage (mixed-methods)

    Dipankar Srirag
    PhD Candidate
    AI & Healthcare Innovation — sequential triage reasoning and bias in LLMs
  • Cogninet Australia

    Industry partner on the CSIRO iPhD program supporting TRIBOT's AI-safety, hallucination-mitigation, and clinical-safety-monitoring workstreams.
Return to top

Publications and media

A curated record of TRIBOT's peer-reviewed publications, conference outputs, preprints, technical benchmarks, and media coverage. This page is updated regularly as the program publishes new work.

  • Coming soon
    • Srirag D., et al. (under review, NeurIPS 2026) — Evaluating Anchoring Bias and Evidence Insensitivity in LLMs for Sequential Clinical Triage: Introducing the TriageSim Benchmark.
  • ACTRI-Bench

    Australasian Clinical Triage benchmark suite for evaluating LLM-based triage agents against ATS gold standards.
    CASMF-Triage

    Clinical AI Safety Monitoring Framework for conversational triage applications — public framework release in preparation.
    TriageSim

    Sequential triage simulation benchmark for probing reasoning biases in clinical LLMs (NeurIPS 2026 submission).
    • NHMRC IDEAS Grant (CIA-A: Narasimhan) — Multilingual Conversational AI for Emergency Triage. Funded program of work, ~A$980K.

  • Coming soon

  • Coming soon

  • Coming soon

Media & Speaking Enquiries

For interview requests, embargoed briefings, or invitations to speak, please use the contact form below and select "Research & Clinical Collaboration" or "Investment & Partnerships" as the reason.
Return to top

Contact us

Partner with use to transform emergency care.

Whether you're a researcher, clinician, investor, journalist, or student, we would love to hear from you. 

For Investors & Strategic Partners

TRIBOT is an NHMRC IDEAS Grant-funded innovation, supported by UNSW Sydney, SWSLHD (Bankstown-Lidcombe), Westmead Hospital, and Prince of Wales Hospital, with industry partnership through Cogninet Australia and international collaboration through TRIBOT-GLOBAL (EPFL, ETH Zurich, CHUV, University of Bern).