The simulations can give you small steps, such as suggesting chemical alterations to your molecule that might have higher binding energy to a docking site that you're targeting, but that assumes that the docking site is stable, that there isn't nonspecific binding, and a whole host of other things, so the first thing you do with those possible changes is run them in a biological system, not a computer.

No one who knows what they're talking about would take the idea that you could do drug discovery entirely in a computer seriously at this point in time.