that's a loaded question. usually you want to stay in the same gamut as your original footage so that you don't get posterization(little lines from rounding errors) if you have lots of different gamuts, you pick from the biggest one being imported.
working in 16 bit is better than 8 bit for the same reason. 32 bit is only needed if you plan on lots of precomps or effects and don't want to accidentally clip values at 255. it drastically changes ways in which alpha channels are used if you leave levels-clipping off. also, a lot of effects don't support 32bpc, so you may get rounding errors. you can use compander/expander to reduce this.
if you set working space to none, you are in essence, using the default color management of your computer monitor. this still won't guarantee correct values because, most likely, the person who created the original footage did not use the same default monitor calibration you did. That's why it's a good idea to always use the guaranteed AE conversion.
As for color finesse, it doesn't natively support color management, but you can burn it in with a precomped utility-profile color converter, then add cf to the precomp. That way it's interface will truly represent your output when you click ok.
http://technicolorsoftware.hostzi.com/
