Accurate modeling of material failure in polycrystalline metals requires an approach that captures both microstructural deformation and damage evolution under complex loading conditions. In this work, we integrate the generalized incremental stress-state dependent model (GISSMO) [1] into a Fast Fourier Transform (FFT)-based crystal plasticity solver from FFTMAD [2] to predict fracture surfaces in cases where experimental testing is unable to provide reliable values due to impractical load-states. To achieve this, modifications to the original FFT solver are done to include heat generation from plastic deformation and thermal softening. Additionally, regulation for the damage parameters of GISSMO are also implemented to avoid localization. The use of this new approach is highlighted for use cases such as ballistic impacts. In these impacts the projectile and impacted surface undergo high strain rates, quickly evolving load states, and damage in the forms such as of dislocations, void growth, spallation, etc.. By having calibrated failure parameters as these develop, the accuracy of failure predictions will improve. This in turn enables improved body and vehicle armour design.