CubicA works like a search engine which doesn’t require internet connectivity, whereby any farmer can simply call a free hotline to get weather and agricultural information anywhere, anytime, in their local language, on any mobile device. Through the integration of machine learning and big data, CubicA leverages previous farmer-caller data (customer journey, location, agronomic context as well as their profile including gender) along with satellite images, weather forecasts and other data sources to guarantee to provide that the right information is delivered at the right time to the right farmer.

The project team successfully completed 4 key phases: 1. Scoping (November 2018) We started with field-level human-centered design workshops and surveys.The purpose of this phase was to (i) better understand Ugandan farmers’ needs, especially in terms of information access, (ii) collect direct feedback from farmers who had never used an IVR (interactive voice response) system like Viamo 3-2-1 Service, (iii) gather direct feedback from active users of the 3-2-1 Service. During this phase, we confirmed our assumption that farmers are willing and eager to access more information through such a system in a more dynamic and tailored way. 2. Design & development (January to June) Based on the insights from phase one, we designed the recommender system that would be deployed in the backend of the existing system (i.e., the 3-2-1 Service). We built the backend infrastructure and database centralizing the different data sources and we developed the recommender system adapting existing open-source models and libraries; this phase was the cornerstone of the project’s technical aspect. 3. Deployment (June to August) During this phase, we prepared the implementation of the recommender system in the backend of the Viamo Uganda’s 3-2-1 Service, drawing from the expertise, knowledge and skills of the Dalberg-Viamo-Bioversity consortium. This project phase required a strong technical collaboration between the teams at Viamo and Dalberg Data Insights, since the former had to integrate the output of the recommender system developed by Dalberg in order to provide a more personalized and dynamic content to 3-2-1 users. 4. Testing & Evaluation (September) Lastly, we tested CubicA by “pushing” the solution to 3000 farmers in Uganda. The testing phase consisted of three rounds of integration of the output of the recommender system in the backend of Viamo’s 3-2-1 Service. During each of the three rounds of testing CubicA was accessed by 1,000 3-2-1 users. We then evaluated our solution by comparing the usage of CubicA with the usage of the standard Viamo 3-2-1 Service. The team looked at the user-journeys of the 3000 people that accessed CubicA with the new recommended messages and compared them with the user-journeys of those that accessed the simple Viamo 3-2-1 Service. This phase confirmed that CubicA is significantly adding value to the system by giving access to more relevant information in a more user-friendly way. Specifically, metrics like the average number of key messages listened to by the users, the average number of calls, and the average number of days in which they called the service, all registered statistically significant positive increases in the 5-10% range.