In ATC's case, the level of engineering expertise applied to drivers, crossover and their integration sort of tells you it's not realistically possible to "better" their design effort without equal test gear, facility and experience. If you change it, it will be "different" which will most likely NOT be better, just different.
I think sometimes the people that talk about driver swapping are overlooking the differences in efficiency of the driver itself in the previous driver vs new driver and overall impact of this issue on driver selection. Small differences of 1/2 dB to 1dB in efficiency in the driver itself applied across the entire segment of the response curve allocated to that driver. Its effectively like turning the entire midrange up or down, or the entire top end (above crossover point) up or down. Even if you could possibly match the response of the driver itself (which is remote), the efficiency issue alone could change the entire character of the speaker. This can have the effect of massive amount of EQ, changing a speaker's entire sound from "bright" to one that's "dark" instantly, or bass heavy to bass light, or mid forward to mid back.
Brad