AI targeting aid after a disaster is faster and often more accurate than traditional assessments. However, it is in and of itself incomplete - people in need will be missed. Can this technology work in concert with traditional on the ground approaches to respond to climate disasters?