Why don't voice radio communications use FM?

Jose wrote:
One proven way to reduce noise is to repeat the signal N times and
the receiver adds up the repetitions and eventually the noise
averages to zero while the signal does not. Of course this is not
what is done in practice in real communications.

Actually, doesn't more transmitter power effectively accomplish that?
(the signal is essentially repeated N times, at the same time)

The unstated assumption is of course that the transmitter is running at
full power output. Once you've done that, the question is what other tricks
are there to increase the range the signals can be detected.