Paper Title
FROM MODELS TO MOBILES TO WHATSAPP: STATISTICS AND AI REACHING THE LAST MILE FOR OILSEED FARMERS

Abstract
Oilseeds occupy a strategic place in India's agricultural economy, underpinning the nation's edible oil security and sustaining the livelihoods of millions of smallholder farmers who cultivate them largely under rainfed conditions. India's edible oil import bill of nearly USD 18 billion in 2024-25 underscores the strategic urgency of the oilseeds sector. The ICAR-Indian Institute of Oilseeds Research (ICAR-IIOR) has served as a leading national institute for strengthening this sector. It is working consistently to translate frontier science into field-level solutions for its six mandate crops (castor, sunflower, safflower, sesame, niger, and linseed) while responding to the persistent realities of climate variability, weather-driven disease outbreaks, and limited access to timely expert guidance in remote regions. Integrating advanced applied statistics with the transformative capabilities of IoT, AI/ML, and Generative AI, ICAR-IIOR proactively translates cutting-edge science into direct, field-level applications that solve critical challenges in oilseed agriculture. This keynote outlines the translational journey from predictive modeling to actionable advisory, illustrating four critical milestones in delivering real-time, on-demand support directly to farmers. The first milestone is a mobile application for the Castor gray mold expert system. It works offline and is built in the vernacular language (Telugu) to serve the farmers of Telangana, where the disease is rampant and causes 100% yield loss. The second is a decision support system sending SMS alerts to castor-growing farmers. Using IoT-enabled Wireless Sensor Networks (WSN) in the farmers' fields, this system periodically sends messages about weather conditions and corresponding advisories. The third is the Oilseeds Crop Doctor prototype, an offline mobile app powered by backend deep learning models that provides advisories based on pest or disease images. The fourth is Oilseeds Kisaan Mitra, a multilingual, AI-powered WhatsApp chatbot which acts as a virtual extension officer. Farmers can message their queries and get immediate answers in their own language. Through an intuitive, natural-language interface, this freely accessible, around-the-clock platform empowers farmers to seamlessly query critical information regarding cultivation practices, hybrid varieties, pest and disease management, and seed availability.Together, these four milestones illustrate how rigorous statistics combined with modern AI can move agricultural science from the laboratory to the farmer's hand, at scale and at speed. Keywords - AI Chatbot, Applied Statistics, Decision Support System, Deep Learning, Farmer Advisory, ICAR-IIOR, Image Analysis, Internet of Things, Machine Learning, Oilseeds, WhatsApp, Wireless Sensor Networks