This is another issue that I have been grappling with. Do you have any word of advice as to how you balance business expectations for a target product delivery vs. the realities of development and the fact that "estimation is hard". It's a conundrum that I frankly don't have an answer for. The reason being, there is a cost-benefit analysis for any product/feature that needs to be developed and it is always a tough conversation when you have to set the right expectations.
The book has a section about the "Iron Cross" of project management: Good, fast, cheap, done. Pick any three you like. You just can't have the fourth. This is also referred to as the Iron Triangle of project management. Agile is supposed to give you data that you can use to make those hard choices.