But not manually!
Automation generally works by automating a manual process. Can you do it manually? What steps do you need to do to get there? Is there any way to do it without being human?
What I mean is, do you have to know language content to find the pauses? If the pauses are almost zero, or if they're the same length as other parts of the show, you're stuck. There's no way to automate that.
This is the same problem that vinyl transfer people have. How to detect the "silent" grooves between the songs. It's not unusual to just give up and do it by hand, because you know content. What's music, what a beginning and end are and what's noise.
There are tools to hack up dialog into matching segments. Put a pause every fifteen seconds and don't pay any attention to the content.