| Summary: | Generative AI (GenAI) technologies offer transformative potential in agriculture, enabling precise and timely advisories tailored to regional needs. The AgroTutor platform developed by CIMMYT pioneers a scalable approach to leveraging localised GenAI to deliver actionable agronomic insights to smallholder farmers across the global south. By employing Retrieval-Augmented Generation (RAG), low-cost language models, and multi-modal data integration, AgroTutor addresses critical challenges such as context localisation, resource optimisation, and quick deployment of RAG. This white paper outlines the process, methodology, pilot outcomes, and the way forward of AgroTutor for agriculture advisories. AgroTutor is an open-source application designed to enable local organisations to provide farmers with cutting-edge, AI-powered advisory services, particularly in the Global South. The system leverages a generative AI framework to mitigate critical knowledge gaps within the agricultural sector. Designed with localised contexts in mind, the system integrates RAG frameworks with multi-modal datasets to enhance advisory relevance and accuracy. Pilots conducted in India, Mexico, and Kenya demonstrate its potential for scalability and ease of deployment, addressing challenges such as crop management, pest control, and climate adaptation.
|
|---|