AgriQuest Voice

Inspiration

Crop identification using satellite observations is an age-old problem but especially difficult in India’s fragmented farmlands. Unlike large homogeneous farms in the USA or Europe, Indian fields are diverse, small, and highly variable. This creates both a challenge and an opportunity: by combining geospatial AI with local engagement, we can build a system that is accurate, scalable, and inclusive for farmers.

What it does

AgriQuest Voice brings together three key layers: Satellite AI Backbone – Uses Sentinel-1 SAR, Sentinel-2 multispectral, and ResourceSAT imagery with the AgriFieldNet India dataset to classify crops across Northern India. Gamified Engagement – Farmers validate AI predictions and adopt sustainable practices through challenges, leaderboards, and rewards. Voice + WhatsApp Accessibility – Farmers interact via local-language voice chat (STT/TTS) or a WhatsApp assistant, ensuring zero learning curve.

How we built it

AI/ML: Temporal CNN + LSTM models fused with radar and optical time-series for robust crop classification. GenAI LLMs: Qwen-3, ChatGPT, and Ollama power multilingual advice, dynamic narratives, and personalized recommendations. Gamification Engine: Challenges (like “Green Cover Badge”) validated by satellite data or IoT inputs. Voice & Chat: Whisper + Azure TTS for voice-first farmers; WhatsApp/Twilio for easy access. Backend & Cloud: Python FastAPI + PostgreSQL + AWS/GCP for scale and reliability.

Challenges we ran into

Handling noisy and incomplete satellite data across diverse regions. Building robust voice recognition for local dialects and low-connectivity environments. Designing gamification that feels rewarding but not overwhelming. Balancing technical complexity with ease of adoption for farmers with varying literacy levels.

Accomplishments that we're proud of

Created a farmer-friendly, voice-first interface accessible in multiple languages. Integrated geospatial AI with gamification for the first time in this context. Ensured scalability by combining cloud infrastructure with lightweight mobile/WhatsApp access. Developed a system that benefits not just farmers but also governments, NGOs, and researchers with accurate crop maps.

What we learned

Farmer-centric design is as important as the AI backbone. Gamification can drive behavioral change when aligned with real-world eco-friendly practices. Voice and WhatsApp channels significantly improve adoption compared to traditional apps. Validation loops (AI + farmer input) improve model accuracy in fragmented farming landscapes.

What's next for AgriQuest Voice

AR Overlays: Smartphone camera guidance for crop health. Carbon Credits: Translate eco-actions into tradable incentives. Government & NGO Integration: Tie rewards directly to subsidies and schemes. IoT & Drones: Enrich monitoring beyond satellites.