It narrows it down though. The release process is a bunch of steps, and errors are either in the process spec or in the execution of it (human error). Using a C

It narrows it down though. The release process is a bunch of steps, and errors are either in the process spec or in the execution of it (human error). Using a CI you remove the possibility of human error, leaving only the option of spec errors - that seems like an improvement to me.