The Gaia dataset will require a huge leap forward in terms of modelling of the Milky Way. Two problems are highlighted here. First, models of the Galactic Bar remain primitive as compared to the Galactic Disk and Stellar Halo. Although Schwarzschild and N-body methods are useful, the future belongs to Made-to-Measure (M2M) models which have significant advantages in terms of storage and flexibility. Second, the Milky Way potential will need much better representation than hitherto. Most models still use very simple building blocks (Miyamoto-Nagai disks or Hernquist bulges) and these will not be fit for purpose in the Gaia Era. Expansions in terms of basis functions offer the possibility of incorporating cosmological information as priors, as well as mych greater adaptability.