For most developing countries, agriculture forms the backbone of food security and economic growth. Data on crop yields, pest and disease control, market and water access, and financial services is crucial to making informed decisions that can sustain agricultural activities. However, traditional data collection methods, such as paper-based government surveys and field visits by extension officers, are often hindered by cost, time constraints, and limitations in remote areas. This study presents an SMS-based chatbot that was built as a solution for collecting agricultural data in these contexts. Unlike common chatbots, which require internet connectivity for them to work, our solution can work offline, making it suitable for our target low-resourced farmers, who usually use feature phones. The full-stack application architecture is described and the technology choices at each level are explained. Early results from functional tests showed that SMS chatbots are feasible, cost-effective, and more efficient than traditional means of collecting agricultural data. The SMS chatbot, which is an AI conversational agent, was also shown to intelligently advise farmers on any agricultural-related topic asked. Lastly, improvements that should be made were noted as further study. As researchers, we are convinced the presented chatbot can use the collected data with little improvements to help key stakeholders make informed decisions, leading to targeted interventions and efficient resource allocation.