ArogyaAI

Breaking Language Barriers in Healthcare

Comment

Inspiration

In rural India, 65% of the population struggles to access timely healthcare due to shortage of doctors, limited awareness, and language barriers. Most health resources are available only in English or Hindi, leaving millions without reliable guidance in their native language. We wanted to create an agentic AI system that combines early healthcare triage with multilingual access to empower rural citizens with trustworthy health information.

What it does

  • Citizens describe symptoms in any Indian language (voice or text).
  • AI translates input → performs preliminary triage → retrieves official health guidelines.
  • Suggests referrals to nearby clinics/telemedicine.
  • Responds back in the user’s native language, with both text and speech.
  • Think of it as a “multilingual AI healthcare companion” that ensures no one is left behind due to language or geography.

How we built it

  • IBM Granite Models + Agent Development Kit (ADK): For orchestrating multi-agent workflows.
  • Translation Layer (IndicTrans / IndicNLP): Handles 22+ Indian languages.
  • Symptom Checker Agent: Uses open medical datasets and fine-tuned LLMs for preliminary triage.
  • Knowledge Retrieval Agent: Fetches information from Ayushman Bharat, MoHFW, and WHO guidelines.
  • Speech-to-Text & Text-to-Speech: Enables seamless voice interaction.
  • Referral Agent: Suggests nearest PHCs/hospitals using open map APIs.

Challenges we ran into

  • Ensuring medical reliability while keeping explanations simple.
  • Building a lightweight multilingual pipeline that works well in rural, low-internet settings.
  • Integrating multiple agents (translation + triage + retrieval + referral) smoothly.
  • Finding open, reliable, and up-to-date healthcare datasets for India.

Accomplishments that we're proud of

  • Built a working prototype that takes symptoms in a local language and returns advice in the same language.
  • Successfully combined agentic AI workflow with multilingual inclusivity.
  • Created a solution that can genuinely improve rural healthcare access and support India’s digital health mission.

What we learned

  • How agentic AI systems can solve real-world problems beyond simple chatbots.
  • The power of multilingual NLP in breaking India’s accessibility barriers.
  • Importance of human-centered design for healthcare tech.

What's next for ArogyaAI

  • Expand to more Indian languages and dialects.
  • Partner with state health departments for real-world deployment.
  • Add offline mode for low-connectivity rural regions.
  • Integrate telemedicine booking and emergency escalation.
  • Explore scaling the same system for legal aid, agriculture advice, and education.
Share this project:

Updates