The AI is a great help. But has a lots of timing and word mistakes depending on the scene. My recomendation is use NUBILES scenes or do not use scenes where there is a lot of moaning and talking at the same time. 

Sorry, I don't understand "use NUBILES scenes". What is Nubiles?

https://nubiles.net/.
Nubiles usually filming scenes where there isnt a lot of moanig and talking at the same time. <span style="color: #ffffff"><span style="font-size: 3%"><span style="font-family: sans-serif">there is a lot of moaning and talking at the same </span></span></span>

Sometimes it's useful to just re-analyse the part of the audio that the AI fail to understand. It can give better result or the same.
Maybe you can play with the options of the AI and try if it's better.

Others AI, not free:
<a href="https://make.headliner.app/login?redirect=%2Fwizard%3Ftype%3Dvideo-transcription">Headliner</a> 10 m free/month account
<a href="https://wave.video/">Wave.video — Easily Edit, Record, Multi-Stream & Host Videos | Wave.video</a> 5m free/month account
<a href="https://www.autosubgen.com/">Auto Subtitle Generator Online (autosubgen.com)</a> Free, but it is no very good.
<a href="https://www.media.io/app">Online Free Video Editor with Simplicity - Media.io</a>