Probably they take batches of 5 minutes of script and ask for bullet points of that, iterating over all the video. Then when they have all the bullet points of

Probably they take batches of 5 minutes of script and ask for bullet points of that, iterating over all the video. Then when they have all the bullet points of 5 minutes packages they can ask for the key bullet points of the total and ask for a conversion to regular text/summary. And for the "see more" that divides all in 5 minutes steps the can ask for a regular text of all the batches