It appears that there are some trade-offs when you try to play a Montage arranger-like.
First, with Montage you can have only eight parts under keyboard control (a fact that is being criticized by Motif users as the Motif series didn't have this limitation). In the YT video five parts are used for the "style" part, three parts remain for solo (right hand) voices. With Genos you have 8 parts for a style, and in addition three right hand parts. Moreover using the registration you can easily switch R1/2/3 voices while performing. To do the same on Montage, you'd use the scene buttons, but these are unavailable as they're already used for switching the "style" variations.
The Yamaha arrangers offer 15 variations (A, B, C, D, 4 fills, break, 3 intros, 3 endings). Montage has 8 scene buttons, so you have to compromise here. Even if you provide one intro/ending only, you loose variations. E.g. with Intro, A, B, C, FillA, FillB, FillC, Ending - 8 scenes are used. In the text it is mentionend that this type of playing applies to both Montage and MODX. It is clearly unusable with a MODX, because MODX has 4 scene buttons only and you must use an extra button to toggle between 1-4 and 5-8. Note that a fill-in will not automatically be followed by the corresponding main variation as you are accustomed. With the Montage you'd have to press a scene button to fire the fill-in and another scene button to return to the main variation.
So these are severe limitations and disadvantages. And there is more, you don't have AI Fingered or the advanced style control that is provided by the CASM section. As an arranger player you obviously want the real thing, and this Montage pseudo-arranger just doesn't deliver, IMHO.
Edit: don't get me wrong, I did own a MOX6, a MOX8, a MOXF8 and now a MODX6 and a MODX7. But it does not replace an arranger.