This talk will be an introduction to my work with Olof Sisask on showing that convolutions are "almost periodic". I will explain how one can show that if f and g are certain functions defined on an abelian group, then f*g has a certain strong translation-invariance property.