Above comments plus:
I'd also suggest different values for R1 and R_load. You're throwing away huge amounts of signal there, and then recovering it in your GAIN opamp (which looks to require a negative supply, is that available?). Connect it as Gunnar suggests, the output will be more linear (assuming you haev a linear pot).
You could use R1 to adjust your output range so that your input op amp with feedback controlled by a pot (output to wiper, one end to inverting input, other end to ground) adjusts your gain appropriately. So for instance, if you want 5 volts maximum, select R1 to give 4 volts maximum or so to the input op amp, then adjust the (greater than one) gain to give you 5 volts.
Another pot & non-inverting op amp to generate an offset voltage. Sum them together with resistor to another non-inverting op amp for your final output. There will be another gain loss doing this, but if you make the resistor applying the offset a fair amount larger than the one applying the signal, this will be minimized, and easily made up for by adding a little more gain. So for example, using 10 kOhms for the signal input and 100 kOhms for the offset input, 11 volts will generate 1 volt of offset in the final output, and the signal will only be attenuated by about 9%.
You could probably combine that last bit into a single op amp. All this might make your adjustments a bit more interactive, so you might need a couple of iterations of adjusting the gain and offset, but the benefit is you don't need a negative supply like your current circuit does, and you save one, maybe two op amps.
Is your 12 volt source regulated? That will have a direct effect on the output.
Oh, and R14 serves no useful purpose (except maybe some SPICE simulators need things like that for convergence).
