Let’s draw a cube. Each vertex is a possible term, AND two adjacent vertices only differ in one variable.
Now, draw a dot for each term from our boolean expression, and group dots that are connected.
An edge that connects two dots means that we can apply the unification theorem to merge those two terms. The variable that differs is dropped.
By applying the unification theorem twice, we can merge 4 vertices that are fully connected.
More generally, by applying the unification theorem 2n times, we can merge n vertices.