Inspiration
I live in Yenagoa, Bayelsa State, Nigeria. I see the problem every day.
A pregnant woman wakes up with a severe headache and swollen feet at 2am. The nearest hospital is 45 minutes away. She doesn't know if this is an emergency or if she can wait until morning. There is no one to call. No nurse on duty. No hotline. Just fear and uncertainty.
This is not a hypothetical. This is daily reality for millions of people across the Niger Delta — one of the most medically underserved regions in West Africa. Bayelsa State has some of Nigeria's worst maternal and child health outcomes. Not because people don't care, but because they don't have access to timely, accurate health guidance when they need it most.
I built BayelsaHealth Navigator to change that.
What It Does
BayelsaHealth Navigator is an AI-powered healthcare triage and navigation agent built specifically for patients and community health workers in rural Bayelsa State, Nigeria.
A patient describes their symptoms — in plain English or Nigerian Pidgin English, exactly how a real person would speak — and the agent:
- Asks 2-3 focused clarifying questions to understand severity
- Triages the situation into URGENT, MODERATE, or SELF-CARE
- Gives specific, actionable next steps tailored to the Niger Delta context
- Recommends the nearest appropriate facility by name
- Understands transportation limitations in rural communities
- Handles common Niger Delta conditions: malaria, typhoid, hypertension, preeclampsia, waterborne diseases, and childhood emergencies
When tested with a real scenario — a pregnant woman with severe headache and swollen feet — the agent immediately identified signs of preeclampsia, classified it as URGENT, directed the family to hospital immediately, warned against unsafe medication, and asked about transportation. That response could save a life.
How I Built It
I am not a developer. I cannot write code. But I had domain knowledge, lived experience, and determination.
BayelsaHealth Navigator is built on the Prompt Opinion platform as a fully A2A-enabled BYO Agent. The architecture includes:
- A2A Protocol: Fully enabled for agent-to-agent communication and interoperability
- FHIR Integration: Connected to AI-Powered FHIR Auto-Fixer MCP with FHIR Context Extension enabled
- MCP Tools: ATLAS Patient Navigator, ATLAS Medication Safety, ATLAS SDOH Screener, ATLAS Clinical Handoff, and CareFlow AI (analyze_symptoms)
- Knowledge Base: Custom VectorStore grounded on a comprehensive Bayelsa State healthcare document covering 13 hospitals, all Primary Health Centres in Yenagoa, disease prevalence data, emergency transport options, and triage quick reference
- Synthetic Patient Data: FHIR-compatible EHR bundle loaded for realistic end-to-end testing
- Multilingual Understanding: Handles Nigerian Pidgin English natively
Challenges I Ran Into
As a non-developer from Yenagoa, every step was a learning curve. Configuring A2A protocols, connecting FHIR MCP servers, building a vector knowledge base, and understanding the technical requirements of the hackathon — all of this was new territory.
But the biggest challenge was also the most motivating: knowing that the problem I was solving is real, urgent, and personal. That kept me going through every error message and every technical hurdle.
What I Learned
You do not need to write code to build something that can save lives. AI agents are powerful tools that anyone with domain knowledge, real-world context, and determination can deploy to solve meaningful problems.
I also learned that the most powerful thing about generative AI is not its technical sophistication — it is its ability to understand context. A rule-based system cannot understand "e dey purge since yesterday" or recognize that swollen feet in a third-trimester pregnancy is a medical emergency. Generative AI can.
What's Next
BayelsaHealth Navigator is not just a hackathon project. It is the foundation of a real product.
Next steps include:
- Partnering with community health workers across Bayelsa State for real-world deployment
- Adding WhatsApp integration so patients can access the agent without a laptop or data plan
- Expanding to cover Ijaw, Itsekiri, and Edo languages
- Building a real-time facility availability database for all hospitals in Bayelsa State
- Scaling to other underserved states in the Niger Delta: Delta, Rivers, and Edo
The goal is simple: every person in the Niger Delta should have access to intelligent, accurate, culturally-sensitive health guidance — regardless of where they live, what language they speak, or how far they are from a hospital.
BayelsaHealth Navigator. Healthcare AI for the Niger Delta. Built from lived experience, for real impact.
Built With
- a2a-protocol
- atlas-health-ai
- english
- fhir
- gemini-flash
- mcp-servers
- nigerian
- pidgin
- prompt-opinion-platform

Log in or sign up for Devpost to join the conversation.