A flat plate is by definition 2 way action. To design it any other way requires large amounts of redistribution, which means that normal estimates of deflections do not apply and crack control is not satisfied unless extra reinforcement is supplied in areas in tension not covered adequately by your one-way reinforcement layout, both in areas with no reinforcement and areas with inadequate reinforcement for crack control. Most design codes limit the amount of redistribution allowed and to design it as one way would violate these limits. From this you would guess that I am opposed to the idea.
I have seen it done before, with strong bands in one direction. But the analysis cannot be based on FEM for the ultimate strength condition as FEM is showing elastic two way action that you are not providing. The analysis has to be based on the design logic being used to be compatible.
For crack control, the analysis should be elastic and this could be based on FEM, to show the areas where an elastic analysis is telling you that reinforcement should be placed.
Then you do not really know what the effect of the redistribution is on deflections.
If done properly you will end up with more reinforcement than you would by designing it properly as a flat slab.