Development in global health is addressed by a complex array of institutions working as “global action networks” (GANs). Network theory suggests a fluidity of connections that is not reflected in most GANs, which are, instead, institutionalized arrangements. We describe the case of a GAN that was ad hoc and temporary. The network successfully produced several now widely used tools for ensuring data quality in systems for monitoring and evaluating programs to reduce the spread of HIV. The ad hoc GAN reflected many of the typical characteristics of GANs, but also exhibited some unique characteristics. Ad hoc GANs focusing on a particular task can be highly adaptive and efficient. We need to learn and foster the circumstances that give rise to them.