IMHO the reason is because engineers can write software but not always solve real problems. It is just easier to put stuff in code to do things with computers. This is the comfort zone. But coming up with something that solves a valuable end-user problem requires understanding what this is.

I've been through this thought process many times and I am still struggling (you can check my profile as to why).

You can try as well.

What are the best applications for AI automation?

People want to build something with the newfound productivity, but it turns out that software development always had a high leverage agains potential impact - coding was always relatively cheap. That means there is in fact no backlog of great products that could have been built if we only had 10x productivity. The only spots where "missing products" can be found are mostly around AI itself.

What is there not to understand? AI is booming right now. People are trying to get in on the gold rush and/or trying to sell pick axes to the miners. There is a ton of fertile ground for anyone that wants to try something, so it's being explored.