A while ago we were looking for tools to automate process documentation and get documentation, SOPs and how-to guides automatically out of the workflow screen recording, but most of the AI SOP generators we tried were too basic and generic (simple click here, click there instructions without any real context), so we ended up building our own tool, video2docs, and really focused on testing different use cases so it could generate context-aware guides instead of simply capturing clicks.
There is still an option to edit the output, because I believe that with any AI tool there should always be a final human review in case something needs to be refined, excluded, or adjusted. And of course, providing good input is important too - all these tools work much better when you create short and focused recordings and split the whole process into smaller processes instead of making one long recording.
It’s similar to giving a colleague a single page of instructions versus handing them an entire book and asking what they need to do based on it - guess in which case you’ll get a clearer and more accurate result.
Maybe this could be helpful for someone else, too.
There is still an option to edit the output, because I believe that with any AI tool there should always be a final human review in case something needs to be refined, excluded, or adjusted. And of course, providing good input is important too - all these tools work much better when you create short and focused recordings and split the whole process into smaller processes instead of making one long recording.
It’s similar to giving a colleague a single page of instructions versus handing them an entire book and asking what they need to do based on it - guess in which case you’ll get a clearer and more accurate result.
Maybe this could be helpful for someone else, too.