Hi All. It's my first post here too, so I apologize if I cover well trodden ground.
I can see two things that the kemper might do to accomplish what you are asking for, but both require jumping through hoops since the Kemper clearly was not designed for this.
1. Linear response correction. This is done by capturing the impulse response of the system in question. This is usually done by exciting a linear system with a known signal, sine sweep, MLS, etc, but that can not be done easily on an acoustic guitar. That doesn't mean that you can't extract an IR though. If you have a pickup on the guitar, you could pluck a series of notes or chords recording the output of the pickup and the output of a microphone. You would then need to do a de-convolution of the dry (pickup) signal and the microphone signal. I believe the resulting IR could be loaded into the cabinet section, and would transform the pickup output into something much closer to the real thing. This technique would be much better than simple EQ as it would include time behavior, which is extremely important in an acoustic guitar due to resonant energy storage in the system (EQ is frequency space only, and does not include time behavior). This is basically a way to capture an accurate EQ with resonant behavior included, but still only deals with the linear response. In a real acoustic guitar non-linear response is what gives the guitar its actual tone, ie. what harmonics are included and in what proportion.
2. Non-linear response correction. This requires actual profiling with the Kemper. I just bought my Kemper, and I have never profiled an amp with it, but I do understand some of the signal theory involved. I think it might be possible to use the built in profiling routines by exciting the guitar body with an audio exciter, such as this one. I have used this technique several times to study the resonant behavior of things like turntable plinths. The essence of what the Kemper is doing during profiling is making a map of the non-linear behavior of the system under test as a function of input frequency and amplitude. I don't see why a similar process could not be applied to the acoustic guitar. I can envision a few problems: One is that you assume that your excitation source, and its coupling to the guitar are linear. Compared to the guitar itself, this might be ok. The other problem is that this information will be duplicated somewhat on your pickup signal. Clearly your pickup already picks up some of the non-linear behavior of the guitar, or it would sound nothing like the guitar at all. If your "profile" of the acoustic guitar was taken with a linear source, but the actual signal it is applied to already includes some of this non-linear response I am not sure where that gets you. Perhaps if you can blend the amp signal with a dry signal it might still get more than the IR only technique. Or maybe it is just a can of worms.
EDIT: I should note that the deconvolution mentioned in the linear response correction section might require some hoops to be jumped through in an of itself. I do not have direct experience with de-convolving arbitrary waveforms like this, but I think it should be possible.
Sorry, I love making things do stuff it wasn't designed to do.
Chad