As mentioned, you can derive trig formulas:
expI(A+B) = exp(IA) exp(IB) = (cosA+IsinA)(cosB+IsinB)
expI(A+B) = cosA*cosB-sinAs(inB + I(cosA*sinB + cosB*sinA)
cos(A+B) = Re { expI(A+B)} = cosA*cosB-sinA*sinB
sin(A+B) = Im { expI(A+B)} = cosA*sinB + cosB*sinA[b]
Note there was no memorization required, just simple laws of exponents. And in many problems you can replace cumbersome trig calculations with simpler more streamlined complex exponential trig calculations.
The complex form of Fourier series is much more compact than the real form and transitions neatly to the Fourier transform.
Whether it is "useful" depends on what you're doing. You won't get far into a signals and systems textbook without Euler's relation. It also shows up in differential equations. There is also a whole field of complex analysis which yields tools such as conformal transformation to determine electric/magnetic field solutions.
But for single-frequency sinusoidal steady state power systems analysis, I don't think it matters whether you represent a phasor as a complex exponential or a magnitude/angle phasor with some predefined rules of what you can do with such phasors (those rules happen to match the rules of complex exponentials). That would be a matter of preference.
=====================================
Eng-tips forums: The best place on the web for engineering discussions.