I'm working on tooling to automatically edit software demo-meetings into video recaps. (E.g. an automatic 90 second "highlight reel" video from each 40 min software demo).

What are some edge-cases in your software demos (or things you can think of) which could cause this system to struggle?

Here's some we've already worked around:

- What if the content is obscured e.g. by the webcams in Zoom?

- What if the user tabs quickly between different things and back?

- What if a modal covers what the presenter is showing?

Some edge-cases we know of but have yet to solve:

- What if the presenter hands control to the user and they do the demo as the presenter explains it

Interested in any difficult edge-cases (real or imagined) which would cause our automatic highlight-reel creation to fail.