I think a "good pipeline to prod" with sufficient automated tests to ensure nothing is broken is the exception not the rule. Even in places that think/say they have a "good pipeline to prod". It's something that takes a shocking amount of engineering effort to do well, and tons of discipline to maintain.