The initial generated stock or a.i. footage isn't accurate to the transcript and what the video is saying, so it doesn't make sense to include it. Even going in and re-generating it with better more thorough descriptions doesn't prompt accurate b-roll footage for either stock or a.i. footage.