I agree with Justin that a common approach is an electron neutralizer. When dealing with insulators however there is a charge balancing act to perform in order to achieve zero net charge. So it may be that you have not enough or too many electrons to compensate for the positive charge on the sample . You may have to fine tune the emission control on your neutralizer to achieve zero net charge for any given x-ray flux on the sample.
In addition to what was suggested earlier you can also try to make your sample as thin as possible and have it on a conducting substrate. Many insulating materials can be studied by XPS if the layer thickness is less than 10 nm, since electrons excited in the substrate can compensate the positive charge at the surface.