In an IFR/IMC environment you have two basic options, proficiency in operations or simplicity in operation. Proficiency requires a good deal of practice. Most GA pilots don't get enough practice on instruments to remain proficient so they the go with simplicity. Simplicity used to be relegated to Autopilot usage. Since most GA planes with an autopilot only have one, this leaves you with a mechanical vulnerability. Constant reliance on an autopilot also has the effect of allowing for degraded hand flying skills and instrment interpretation skills from lack of practice.
You make the options you've suggested out to be an either/or scenario. Either be simple, or be proficient.
Autopilot operation does not simplify the flight. It complicates it to some degree, or at a minimum alters one's handling of the aircraft to include monitoring an additional component, and managing it.
Proficiency is always necessary. An autopilot does not take the place of proficiency at instrument flight, at visual flight, at night flight, at hand flying, or at autopilot operation. It's not a system one can or should simply turn on and go, especially in light aircraft. Even in significantly more advanced systems that enable the autopilot (if properly utilized) to be used for almost the entire flight (including landing), it's far from a substitute for piloting. It's an aid, a system; something to be used while flying the airplane, but at no time does the autopilot or aircraft flight control system become a substitute for piloting the aircraft. It's simply a tool to be used while doing so.
Accordingly, introduction of automation into the cockpit increases the amount of training and proficiency one must accept and undertake on an initial and ongoing basis. It's one more thing at which one must be proficient, and maintain proficiency.
One need not and should not make a choice between proficiency and simplification (in the context as you've introduced it, simplification appears to mean automation). Both are always necessary. One might as easily substitute "management" for "simplification," as an autopilot is both a management tool, and a tool that must be managed.
I use autopilots that have substantially more capability than a typical light airplane autopilot; they have CAT III autoland capability, with autothrottles and navigation through a number of redundant GPS/FMS systems. None of that at any time is a substitute for my own proficiency, nor would it ever be acceptable to utilize the automation in lieu of proficiency or skill. An autopilot is not a crutch.
During training, on a regular basis, while flying an approach the instructor in the simulator will fail the autopilot. One may be conducting a three-engine approach with the autopilot engaged, for example, and will experience additional problems or emergencies that will eliminate the autopilot, or the instructor will simply fail the autopilot during the procedure, if he desires. One should be (and must) always prepared to immediately disconnect and continue hand flying the aircraft while addressing the problem. There need not be a problem; taking over by hand is always an option that should never be discounted. Pilots who abdicate being in command of the aircraft in favor of letting the autopilot do their job aren't experiencing a lack of skill or degradation in their abilities as a result of autopilot use. They've failed to fulfill their basic function for being PIC: they've given up command. The failure then, and the root cause of any decrease in skill or proficiency isn't the autopilot, it's a lazy pilot.