We are currently doing what Rod suggests in a large project of 10 courses. But we increase the height of the pseudo CC box to accommodate more text. There isn't a way to cycle different amounts of text for long audio files.
We actually can interject with JavaScript, text into Captivates CC, but the same problem exists, unless the amount of CC text is hardcoded. Right now this has to be done post/publish for every file. Not a good solution.
Another option is to use one audio file and use micro-navigation to jump around the slide, with click boxes to pause the slide at the end of each section. This way you can use Captivates CC.
We are currently working on a JavaScript solution to set-up the micro-navigation dynamically, so that changes to the slide do not affect the timing of the CC. Also the pausing will happen dynamically without using the click boxes as you can determine how many audio files there are and their length.