Unimesh would make things different. You cannot say in general things would be easier or less easy. 

For an Unimesh figure Indeed all geometry would be one mesh. Currently in Poser a figure is a collection of independent object meshes. The meshes are linked to a bone. The meshes are deformed when the figure is posed. Where two object meshes meet there is a 'weld' and corresponding vertices from both sides are merged in a location that is the average of both positions.  

When a figure is a composition of different object meshes, it is easy to make them show or hide individually, change them individually (replace body part with prop). In the original Poser definition of bend weights, via zones and falloff angles, the degree of participation in the bending was determined by the xyz coordinates of the vertices so as long as the vertices at the 'interface' with other body parts one could swap out geometries at will, irrespective of the vertex count or mesh topology away from the boundaries. Therefore the solution gives optimal ease of grafting. 

Disadvantages of independent objcts deformed-to-look-like-one-surface become apparent when you want to add vertex related information like morphs and mapped bending weights because you have to define them for each individual object mesh. Also difficult is operations and processes that work over a range of polygons/facets, such as sub-surface scatter, subdivision, weight/morph painting. The affected area may extend across two or more objects.

For each of these operations, that have become standard practice now, masks and wrappers were developed to make them feasible on the old core system but they all came with limitations and at a price in terms of performance. 

I therefore expect a truly Unimesh Poser would see a performance boost but a limitation in 'graftability'.