We tackle the role of austenite in multiphase steels on hydrogen diffusion systematically for the first time, considering a range of factors such as morphology, interface kinetics and the additional effect of point traps using both experiments and modelling. This follows the findings from part I where we showed that austenite cannot be parametrised and modelled as point traps under the assumption of local equilibrium, unlike grain boundaries and dislocations. To solve this, we introduce a 2D hydrogen diffusion model accounting for the difference in diffusivities and solubilities between the phases. We first revisit the as-quenched martensite permeation results from part I and show that the extremely low H diffusivity there can be partly explained with the new description of austenite but is partly likely due to quench vacancies. We then also look at the H absorption and desorption rates in a duplex steel as a case study using a combination of simulations and experiments. The rates are shown to depend heavily on austenite morphology and the kinetics of H transition from ferrite to austenite and that an energy barrier is likely associated to this transition. We show that H diffusion through the ferrite matrix and austenite islands proceeds at similar rates and the assumption of negligible concentration gradients in ferrite occasionally applied in the literature is a poor approximation. This approach is also applicable to other austenite-containing steels as well as other multiphase alloys.