-
-
User Interface for Interaction
-
Mobile Interface
-
Multi Agent Flow
-
Regional Conversation for Farmer
-
Farmer Advisory Conversation
Inspiration
Agriculture sustains billions, yet smallholder farmers and local traders often lack access to timely expertise, fair pricing, and data-driven decision-making tools. We wanted to leverage multi-agent AI and RAG to bridge this gap, starting with India as our pilot case, while building a system that can scale globally.
What it does
AgroSahayak is a global multi-agent AI assistant for farmers and traders. It provides personalized, multilingual guidance on crop management, irrigation, pest and disease control, weather-based planning, finance, market pricing, and government policies. Farmers gain actionable insights to boost yields and reduce losses, while traders access real-time demand forecasts, transparent pricing, and optimized supply chains.
How we built it
We designed a multi-agent architecture where each agent specializes in a task—query understanding, retrieval (RAG), weather, crop advisory, irrigation, pest/disease detection, finance & market analysis, policy updates, fact-checking, explainability, and multilingual responses. We integrated data from IoT sensors, satellite imagery, government databases, and real-time commodity markets. A conversational interface powered by generative AI delivers outputs in users’ preferred regional languages.
Challenges we ran into
• Integrating heterogeneous data sources (IoT, satellite, government APIs, market feeds). • Balancing real-time performance with accuracy in RAG-based retrieval. • Designing AI explainability for farmers and traders with minimal technical background. • Multilingual support with regional dialect nuances. • Framing a localized pilot (India) while keeping the solution globally scalable.
Accomplishments that we're proud of
• Built a functioning prototype of a multi-agent RAG system tailored for agriculture. • Created a multilingual conversational interface for inclusivity. • Demonstrated use cases for both farmers (crop yield optimization) and traders (transparent trade). • Positioned the system as globally adaptable while testing on one of the world’s most complex agricultural ecosystems India.
What we learned
• The importance of modular AI agent design for scalability across different geographies. • How explainability builds trust when users rely on AI for critical decisions. • That combining social impact with cutting-edge AI significantly strengthens innovation pitches. • Agricultural intelligence requires constant real-world validation to remain relevant.
What's next for AgroSahayak
• Expand pilots across India to refine models with diverse farming conditions. • Partner with global agri-cooperatives and NGOs to test in Africa and Latin America. • Incorporate computer vision for crop disease detection using smartphone cameras. • Strengthen predictive analytics for climate-resilient farming. • Build APIs for integration with agri-fintech and supply chain platforms.
Log in or sign up for Devpost to join the conversation.