Depending on the system, the simple adsorption models can actually be pretty good. A good discussion of the limitations and a statistical derivation can e.g. be found in Desjonqueres/Spanjaard: Concepts in Surface Physics.
Langmuir: the adsorption is limited to a monolayer or, at least, for low coverages multilayer formation is so unfavorable that it barely contributes. Lateral interaction is negligible.
BET: multilayer formation is substantially contributing and may even go so far to contribute.
Fowler-Frumkin: includes lateral interaction within a monolayer, but depending on the parameters, you get a non-functional loop.
If the adsorption turns more complicated, e.g. when diverse chemisorption kinetics are happening or when the material doesn't allow for clean layer differentiation as it is the case in some porous materials, these isotherm equations may turn too simple.