As the closest Seyfert nucleus, NGC 1068 is a key object to establish on firm grounds the model of active galactic nuclei (AGN). Two important questions relative to AGNs are : i) can we actually see the parsec-scale torus of dust and molecular gas around the central engine predicted by the popular Unified Model model (e.g. Antonucci, 1993)?; ii) how the gas is brought down to the parsec scale to feed the monster, and, in particular, is there a micro spiral/bar system that could produce the required torque ? Adaptive Optics in the near-IR has potential advantages to make a significant progress in those two questions : the low dust opacity and the excellent angular resolution.