HealthAI: Intelligent Healthcare Assistant Using IBM Granite

PROJECT TITLE : HealthAI: Intelligent Healthcare Assistant

Using IBM Granite

TEAM SIZE : 4

Team Leader : Gangireddy Mahendra Reddy

Team member : Bathula Vishnu

Team member : Bala Harini Palnati

Team member : Deena Mannepalli

Team ID : LTVIP2025TMID37458

College : Rise Krishna Sai Prakasam Group of

Institutions

VIRTUAL INTERSHIP PROGRAM

Intership details : Generative AI with IBM Cloud

Internz : Smart Internz & IBM

Company : Smart Bridge powdered by Smart Internz

📘 Introduction

Project Overview
HealthAI is a real-time intelligent healthcare assistant built using Python, Streamlit, and OpenAI/IBM
Granite. It simulates doctor-like responses, predicts diseases, suggests treatments, and visualizes health
metrics. It demonstrates how AI can be leveraged to provide meaningful health advice with a simple UI.
Purpose
To leverage IBM's Granite LLM for building a reliable, accessible, and AI-powered health assistant that
aids users in health-related queries efficiently and securely.

💡Ideation Phase

Problem Statement:
Many individuals struggle to access quick and reliable health guidance, especially in remote or
underserved areas. HealthAI bridges this gap using AI-powered assistants.
Empathy Map Canvas:
Think & Feel: Wants accurate medical advice
See: Too much online misinformation
Say & Do: Seeks help through apps
Hear: Concerns about trust in AI
Pain: Long wait times at clinics
Gain: Fast and trusted guidance
Brainstorming:
Use AI/ML for health Q&A
Integration with IBM Granite
Chat interface for ease of access

Data Flow Diagram:

1|Page Team ID: LTVIP2025TMID37458

✅Reqirement Analysis

Customer Journey Map:

User opens app → asks question → receives response → gets suggestions

Solution Requirements:
• AI model: IBM Granite
• Backend: Python, Flask/FastAPI
• Frontend: HTML/CSS/JS or Streamlit
• Database (optional)

📜 Declaration
We hereby declare that the project titled "HealthAI: Intelligent Healthcare Assistant Using
OpenAI" is the result of our original work completed under the guidance of SmartInternz
and IBM. It has not been submitted elsewhere for any academic or professional recognition.
All development, documentation, and implementation tasks were carried out with sincerity
and adherence to academic integrity.

🙏 Acknowledgment
We express our heartfelt gratitude to SmartInternz and IBM for providing us this
opportunity to work on an industry-relevant guided project. We also thank our faculty
members and mentors for their continuous support, technical guidance, and valuable
feedback throughout the course of this project.

🧠 Abstract
HealthAI is a real-time intelligent healthcare assistant that enables users to interact with an
AI system to receive health-related information. Built with Python, Streamlit, and OpenAI,
the system offers personalized responses to health queries, provides AI-powered disease
predictions based on symptoms, suggests treatment guidelines, and visualizes patient data
through insightful graphs. It serves as an educational and awareness tool, empowering
users to make informed decisions about their health.

2|Page Team ID: LTVIP2025TMID37458

3|Page Team ID: LTVIP2025TMID37458
Ị Problem Statement
Access to timely and professional medical advice remains a challenge for many, especially in
rural and underserved communities. HealthAI aims to bridge this gap by offering a virtual
assistant capable of providing preliminary health guidance, improving awareness, and
reducing dependence on overburdened healthcare systems. It acts as a first point of
reference before seeking clinical care.

🛠 Tools and Technologies Used

Tool/Technology Purpose

Python Backend logic and data processing

Streamlit Web interface development and deployment

OpenAI (ChatGPT) AI-based natural language understanding and response generation

Pandas Health data manipulation and storage

Plotly Health metric visualization through dynamic graphs

GitHub Version control and collaborative development

SmartInternz Platform Project structure, guidance, and submission support

🧩Modules Implemented
Patient Chat – Offers interactive, AI-generated responses to user health queries using
prompt engineering and natural language processing.

Disease Prediction – Accepts symptoms from users and predicts possible conditions using
symptom-based AI analysis.

Treatment Plans – Provides standard medical guidance and lifestyle suggestions for known
medical conditions.

Health Analytics – Visualizes user health data trends like blood pressure, glucose levels, and
heart rate, with AI-generated insights for better health tracking.

4|Page Team ID: LTVIP2025TMID37458

⚙️How to Use the HealthAI Application
1. Open App:HealthAI Streamlit App
( https://healthaiproject-9zncftohtgbtxnwcvrkdss.streamlit.app/ )

2. Choose Feature: Patient Chat / Disease Prediction / Treatment Plans / Health

Analytics
3. Input Data: Type symptoms or condition
4. Submit: Click generate
5. Output: Get response from AI instantly

Sample Module Outputs

·.t_*± Home Page:

The HealthAI homepage provides an intuitive interface with quick access to modules like
Patient Chat, Disease Prediction, Treatment Plans, and Health Analytics. It displays patient
details and allows users to begin interacting with the AI assistant immediately.

·¡_^ HealthAI: Intelligent Healthcare Assistant

Welcome to HealthAI, your virtual healthcare companion powered by AI.

This application provides:

• 'Real- time medical chat support using AI

˘ ⬛ _*’

• C˛* Disease predictions based on symptoms

• ˙’Personalized treatment plan suggestions

• /¡ç# Health data analytics and trends

5|Page Team ID: LTVIP2025TMID37458

< Module 1: Patient Chat

User: I have fever and body aches.What could it be?

HealthAI :

• Malaria: A serious disease spread by mosquitoes causing fever and flu-like

illness.
• Chikungunya: A viral disease transmitted to humans by infected mosquitoes.
• Pneumonia: An infection of the lungs causing inflammation and fluid buildup.
This is not a diagnosis

`^ Module 2: Disease Prediction

6|Page Team ID: LTVIP2025TMID37458

User: fever,cough,rash,body pain

HealthAI :

Analyze Symptoms
Possible conditions based on your symptoms:

• Chikungunya
• Pneumonia
• Tuberculosis
• Cold
• Dengue

˙’Module 3: Treatment Plans

Treatment Recommendations
Enter a medical condition:
• Malaria
Treatment Options for Malaria
• Medications:
• antimalarial drugs
Lifestyle Recommendations:
• use mosquito nets
• avoid stagnant water

7|Page Team ID: LTVIP2025TMID37458

# Module 4: Health Analytics
/¡ç
Displays 30-day trends of Heart Rate, BP, Glucose with insights.

🏗 Technical Architecture

Proposed Solution:
Interactive chatbot powered by IBM Granite that processes symptoms and suggests outcomes

8|Page Team ID: LTVIP2025TMID37458

📅 Project Planning & Scheduling
Day 1: Research&Planning
Day 2: Backend & AI Integration
Day 3: Frontend Design
Day 4: Testing & Debugging
Day 5: Documentation

🔁 Project Workflow
- Requirement Gathering
- AI Model Selection
- UI/UX Planning
- Development (Frontend + Backend)
- Integration & Testing
- Deployment

🚩 Milestone 1: Model Selection and Architecture

We selected OpenAI GPT-3.5 for its powerful conversational abilities. Backend and frontend
were designed to interact through clean APIs and modular components.

🔧 Milestone 2: Core Functionalities Development

Each module was developed independently:
- AI chat for general queries
- Symptom-based disease prediction
- Condition-based treatment planning
- Time-based health data visualization

📜 App.py Development
This main file handles routing, input collection, response display, and connects user
interactions to utility logic and the OpenAI model.

🎨 UI Design
User interface is built using Streamlit’s built-in widgets and layout elements. Simple yet
modern layout using sidebar, form inputs, charts, and response sections.

🚀 Deployment
Hosted on Streamlit Cloud. Environment variables securely stored. GitHub for code
management and team collaboration.

9|Page Team ID: LTVIP2025TMID37458

🔮 Future Scope
-Connect with verified doctors
-Use wearable IoT devices
-Add user login and history
-Deploy on cloud with authentication

✅Advantages & Limitations

Advantages:
Instant AI responses
Easy-to-use interface
Based on reliable IBM/OpenAI models

Limitations:
No real-time doctor verification
Requires internet
Accuracy depends on model training

✅ Conclusion
HealthAI was a hands-on experience in building AI-integrated health apps. We learned
prompt engineering, Streamlit deployment, and OpenAI usage for real-world healthcare
problems.

🔗 Project Links
App: https://healthaiproject-9zncftohtgbtxnwcvrkdss.streamlit.app/
_¿□ GitHub: https://github.com/mahendra4338/HealthAI_Project/tree/main
DemoVideo: https://youtu.be/nyhc0tBoAZQ?si=tzA0AElV9AcZ2XQo
Report Source: IBM Granite Model Documentation

10 | P a g Team ID: LTVIP2025TMID37458

e
11 | P a g Team ID: LTVIP2025TMID37458
e