Why are there so many failed attempts in attended automation? Automation, although it sounds easy, it's actually very difficult, especially when you have a person involved in the process as well. Different people have different ways of doing things, even if it's the same thing. Different people would have different understanding of what is done for them automatically. Because of not only dealing with the little quaint things that the applications themself have, you also have to deal with how the agents syncing together with the bot will behave. And you have to account for errors and mistakes that could happen due to that or different kinds of exceptions. The idea with attended automation is that you have to think, when you build the automation, you have to think how would the person work together with the automation and that is mostly a difficult thing. So you can't just wait on things to happen assuming that they will happen. You have to handle the cases of what happens when they don't happen the way that you perceive them?