Artificial Intelligence-Based Chatbot Model Providing Expert Advice to Potato Farmers in Kenya
Meshack Kiplangat Korir, Waweru R. Mwangi, Michael Waema Kimwele · 2023
This research paper presents a solution to the declining yield in potato farming in Kenya, which has been attributed to a limited supply of quality seeds and access to expert advice. The study used IBM Watson Assistant, an AI-based chatbot framework, and GIS to offer expert advice and link farmers to quality seed producers. The paper first introduces potato farming in Kenya and provides a brief history of conversation agents before delving into the theory behind chatbots, including their classifications and general architecture. The methodology section outlines the five significant steps taken in the research, including data collection, implementation, testing and training, and evaluation. The evaluation phase used performance indicators which are presented in detail. The results demonstrate that this potato farming chatbot model had a score of 97.7% in terms of message coverage, a score of 78.4% in terms of conversation containment, and was 88.05% effective, users were 60% satisfied with the model and the likelihood of use of the model in the future was at 80%. The study concludes that this integrated potato farming chatbot model is a practical solution for farmers to improve their yield, and the recommendations made based on user feedback and expert input could improve the model further. Overall, the study presents a promising location-based approach to addressing food security challenges in Kenya through technology-driven solutions.