If you turn the model inside out (apply a linear transform with a matrix that has negative determinant) then the surface normals will be reversed, too. Using Surface Toolbox module is a good solution. Is there any specific reason you would only want to turn the surface inside out but not flip the normals?