FMVs allows you to do things that the engine cannot. Mode 7 could not show a character's face up close. Mode 7 could not show an animated city, only pieces of it, a finite number of tiles at a time. Even when Mode 7 showed far away shots of Narshe, it couldn't then come into the controllable overhead view of the player. These three methods show how the first two minutes of Final Fantasy VII alone had more sophisticated techniques than the entirety of FFVI.
The sequence I just described was the first time a
sequence shot had ever happened in an interactive medium. A more recent example would be the Metal Gear Solid: Ground Zeroes trailer. The sequence shot is an incredibly difficult and coveted technique in film that takes a deal of creativity and skill to pull off. The fact that its use also serves as a nod to the film industry increases the creative ingenuity of Square using it.
I can't imagine how allowing creators express themselves in different ways is not an important thing.