I would not rewrite every auto-generated chapter just to make the labels sound polished. The useful question is whether a viewer can reach the section they need without scrubbing through the video. Review the draft against three things: does each label describe the section it opens, is there a missing chapter where the viewer’s task changes, and would fixing either problem make navigation noticeably easier? A dull label is cheap to leave alone. A label that promises the answer but points to two minutes of setup is worth correcting. Chapter polish has little value if it does not help someone use the video.
Consider an 18-minute tutorial that moves from choosing equipment to configuring it, then spends four minutes troubleshooting a failed setup. If the draft groups configuration and troubleshooting under “Setup,” a separate troubleshooting chapter gives viewers a useful place to jump to or return to. A brief aside during configuration probably does not need its own chapter, even if it has a neat label. That is the editing tradeoff: fix chapters that help viewers find a substantial payoff, and leave harmless imperfections alone. Where do you draw that line—what label or missing transition has made you revise an otherwise usable chapter draft?