AI Audio + Pronunciation + Correct Closed Captions

I'm having an issue with the new AI Audio feature in Storyline, as it relates to pronunciation and generating CCs.

In the past, using Text-to-Speech, I often needed the voice to pronounce something differently than how the word appeared. 

I'd generate and insert the voiceover with "Generate Closed Captions" checked: "Welcome to Brivo."

Then I'd go back into that voiceover, change the words to phonetic spelling according to how I needed them pronounced, uncheck "Generate Closed Captions," and insert: "Welcome to Breevo."

This would result in correct CCs and correct pronunciation. The audio would state "Welcome to Breevo" while the CCs would read "Welcome to Brivo."

I'm unable to duplicate this workflow using AI Audio. When I uncheck "generate CCs," it simply does not insert any CCs at all. I'm left having to waste so much time generating and re-generating audio in hopes that it randomly pronounces all my words correctly.

Has anyone figured out a workaround for this issue? Am I just not using the tool correctly?