Digital connectivity is essential for advancing UN SDG4: quality education. Accurate and complete information on the locations of schools and their connectivity status is crucial for identifying gaps in infrastructure. In this work, we introduce a novel AI-enabled school mapping and internet connectivity status prediction workflow, in support of digital capacity building. First, we investigate the performance of state-of-the-art computer vision models for school mapping. Next, we introduce a connectivity prediction model using machine learning and publicly available remote sensing data. We evaluate our approach in five pilot countries: Bosnia and Herzegovina, Belize, Botswana, Guinea, and Rwanda. Finally, as a proof-of-concept, we run our pipeline end-to-end in 10 districts in Botswana.