The analysis of how much time does an automation save vs how much time to build is important. I'd like to see more instances of that being used and it would be really nice if we would add 3 factors to it.
What is the error tolerance for the manual activity? If a process only takes 4 seconds to complete and is done once a week and an automation will take 1 week+ to build it hardly seems worth automating; that is of course unless the data produced in that 4 seconds is so critical that an error in it comes with disastrous consequences. At that point automation goes from "meh", to mandatory. Even if the consequences aren't quite disastrous and are simply impactful or even just annoying it can be worth it and warrants evaluation.
What happens if it isn't done? A quick manual process is easy to forget doing especially by a human so the impact of the task not getting done may be enough to make automation worth it.
Optics (aka perception, marketability). Sure a process is simple and it takes only a few seconds but if making a user do it leaves them with the feeling, "It's 2022, I can't believe I still have to click that stupid button." then automation will have more value than just time saved.