AI-Based Farmer Query Support & Advisory System(AgriSaarthi)

🌱 Inspiration

Farmers often need quick and reliable answers about crop diseases, pests, fertilizers, irrigation, soil management, market-related issues, and government schemes. However, agricultural extension services can face a high volume of farmer queries, making it difficult to provide timely assistance for every problem.

We wanted to build a system that could provide fast AI-based assistance for routine agricultural questions while ensuring that complex cases reach the appropriate agricultural expert.

This inspired us to create AgriSaarthi – AI-Based Farmer Query Support and Advisory System.

💡 What is AgriSaarthi?

AgriSaarthi is a full-stack intelligent farmer advisory platform that combines AI, NLP, and expert assistance to help farmers get faster and more relevant support.

Farmers can submit their queries through text or voice, with support for multiple languages. The system analyzes the query and identifies the agricultural domain, crop, problem, complexity, and confidence level.

Based on the analysis:

Simple and reliable queries can receive an AI-generated advisory response.
Complex, ambiguous, specialized, or low-confidence queries are escalated to an agricultural expert.
The system helps route expert cases according to the relevant agricultural domain.
Farmers can track their query status and view their previous queries and responses.
Experts can view assigned cases, analyze farmer queries, and provide expert advisory responses.
Administrators can manage users, experts, and the overall advisory workflow.

🛠️ How We Built It

AgriSaarthi was developed as a full-stack application using:

React.js for the frontend
Node.js and Express.js for the backend
Python for AI/NLP processing
MySQL for data storage
OpenAI API for AI-powered query understanding and response generation
REST APIs for communication between the frontend, backend, and AI services

The AI processing pipeline analyzes each farmer query and produces structured information such as:

Domain → Crop → Problem → Complexity → Confidence → Expert Requirement

For example, a straightforward question about preparing organic fertilizer can be handled by the AI advisory system. A query describing widespread crop damage or an uncertain disease condition can instead be escalated to an expert.

🚜 Key Features

Farmer

Text-based query submission
Voice query support
Image upload for agricultural problems
AI-powered agricultural advisory
Query history
Query status tracking
Multilingual interface
Expert escalation for complex cases

Expert

Assigned case management
Agricultural domain-based case assignment
Farmer query and uploaded media
AI-generated analysis and escalation reason
Expert advisory response
Case status tracking

Admin

User management
Expert management
Expert assignment
Query monitoring
Advisory workflow management
Knowledge-base management

📚 What We Learned

Through this project, we learned how to integrate Generative AI with a real-world full-stack application rather than using AI as a standalone chatbot.

We gained practical experience in:

Designing AI-powered workflows
NLP-based query classification
Prompt engineering
REST API integration
React frontend development
Node.js/Express backend development
MySQL database design
Role-based authentication and authorization
Expert escalation workflows
Multilingual application design
Connecting Python AI services with a web application

One of our main learnings was that an AI system should not attempt to answer every query. Recognizing uncertainty and routing complex cases to human experts is an important part of building a responsible advisory system.

🧩 Challenges We Faced

One of the biggest challenges was designing a workflow that could reliably distinguish between queries that AI could handle and cases that required expert intervention.

We also faced challenges in:

Maintaining consistent AI classification
Preventing incorrect or overconfident agricultural recommendations
Designing domain-based expert assignment
Connecting the React, Node.js, Python, and MySQL components
Maintaining consistent query IDs and statuses across the system
Supporting multilingual user interfaces
Handling text, voice, and image-based farmer queries
Preventing duplicate query and expert-case creation

We addressed these challenges by creating a structured query-processing workflow with clear statuses and explicit escalation conditions.

🌾 Future Scope

AgriSaarthi can be extended with:

More regional Indian languages
Crop disease image detection
Weather-based recommendations
Real-time market price integration
Government scheme eligibility assistance
Personalized crop recommendations
Advanced agricultural prediction models
Regional agricultural knowledge bases
Mobile application support
Analytics for identifying common farmer problems and knowledge gaps

🎯 Our Vision

Our vision is to make agricultural assistance faster, more accessible, and easier to reach, while combining the speed of AI with the experience of human agricultural experts.

AgriSaarthi aims to act as a digital bridge between farmers, AI-powered agricultural knowledge, and expert assistance.