System failure and how do we recover?

Interesting problem – when a system fails how do you recover?
Human intervention?
Most system failures if they are systemtic would require human intervention, but when do you do that?

Recently I have been interacting with a company that is famous for it’s automated system. Only to find out the automated system is only a tier 1 response. Tier 2 would in fact require a human to intervene. Now, because of customer service is that in fact the right thing to do? Listening to someone pound their keyboard into dust as they are trying to help me isn’t effective for me. It may be for other people but not for me.

So when should your system be automated in recovery and when should a human being intercede?

It’s an interesting developmeant and architectural question.