How difficult is it to sync audio playback to cgi lips or mouths?

It is not difficult at all since the vocal performance is always recorded prior to animating the CGI sequence. The vocal animator is given a timeline segment containing the vocal performance and the character performance, they simply insert the appropriate mouth shapes (including chin and tongue movements) for each syllable at the appropriate key frames. The animation software then fills in the "tweens" from one mouth position to the next which the animator can fine-tune by inserting additional key positions where appropriate. Once the sequence is playing smoothly, the sequence can be passed to the compositor.