1: ya, it’s easiest to have animation drawn to music. one way i worked in the past was come up with a skeleton track for the animator - something with a basic beat and outlying minimal changes to song-structure - and because they had the tempo overall from that, we could change the music from there easily - i also worked with a student on an independent film where they wanted me to follow their visual cues(because sometimes the music needed to be rhythmic, other times not, and they wanted me to keep with the rhythm of their overall visual story), it can go either way, but it’s easiest to give the animator something to draw their own visual rhythm to, early on.
Generally speaking, though, the animator should probably be the one to come up with a simple ‘storyboard’ first, so that you both have a narrative in place ahead of time that you can plan around(it depends on how complex the story will be; if it’s simple enough to convey by word-of-mouth and remember, then it’s fine, but for longer narratives, it can be very necessary for smooth collaboration… otherwise, you’ll run into many situations where you’ll both clash with each other saying, “i thought when you said [that] you meant we should do [this]” 
).
2: no advice from me which is Ableton specific… just for any DAW, pay special attention to saving presets - you and the animator want to achieve a uniform look and sound across sections, but to get nicely detailed where needed, you may need to break things up in sections and work modularly - presets help you dial everything in automatically no matter how much you edit timings in different sections(some visual artists also use templates for themselves, where they create one finalized frame of animation early-on to refer to for the look and coloring of everything else).
3: ya, to some degree you’ll always end up working modularly(different parts of a film require differing levels of detail and attention - music videos are least like this because the song makes it somewhat uniform throughout, but for films with narratives, the feel can change more often; all different contexts are defined by so many different variables… for example, you could look at ‘Silly Symphonies’ at the 6:20 mark, compare how you would do that level of action and complexity to something simpler like when she comes down the steps around 3:30 - you would probably want to work on the 6:20 action more in detail, over a greater amount of time, and allow greater flexibility for the artist to change things around there often). Certain scenes are much more important and climactic, and will require more time to get it just right, so you’ll work longer and harder on it as you’ll feel the audience-reaction will hinge on that particular scene more than any other.
Have fun! Best of luck 
