Your assessment is good and correct, because AI agents or models should only be used as assistants in doing anything, and the person who is using them as the assistant should know everything about what they are doing, so that if the model or agent makes a mistake, they should be able to understand that it's a mistake and should be able to rectify it.

Someone who doesn't know what they are doing, and believes that the AI is going to do everything for them and they just need to sit back, and the agent will keep making them money consistently, is doing it the wrong way, and they will most probably lose more money than they will make using such agents and models.

People need to understand that AI models are only good if they are used for help, and you can't rely on them for everything, or expect them to do everything correctly without any human input, that's basically not possible, at least not for now.