On the plus side - I encountered a guy who was whinging about questions on a CAD blog. The subject was how people converted from fully-machined parts to cast parts and he thought that was the most ridiculous thing in the world.
I mentioned that not only did that happen, but I had personally seen a weldment design discarded due to the cost of the individual x-raying of each segment for a casting design that was then discarded for lead time to a hogout which was then discarded due to machining cost with the people in charge saying "Why isn't this a weldment?" The same people who discarded the weldment in the first place.
After he recovered from that he related that where he worked when a new design was investigated the entire company process was represented: design engineers, stress, materials, manufacturing, assembly, inspection, procurement, contracts, and management. They continued meeting as a cohesive group until all were satisfied they all understood their part and that, collectively, the group understood what was going to happen.
That seemed like an ideal way to operate and, while the up-front cost showing up is what I am sure is the reason it's avoided, it certainly means that if any problem occurs then everyone involved has a good idea of the trade-offs that are reasonable to cope without having to start from scratch.
I wonder if he had understood just how lucky he was not to work in a place that regularly played "toss the grenade over the wall."
So, in answer to the original question - that should all happen at the same time and the feasibility study that is supposed to be part of systems engineering, would already have the internal workings identified to correctly rate all the other "ility" components before proposing this as a workable system.
==========
In my case this disconnect was partly due to a forced merger with another division that exacerbated the problem, and particularly that the other division had been so remotely located from their minimum wage factory that the factory, having never been well supported, decided to just do whatever they wanted. For them, engineering was needless overhead. The factory regularly informed engineering of the design changes that were required to document the product they had already produced. If those changes weren't instantly made the factory would report that "Deliveries were going to be delayed because <insert engineer's name> wasn't doing their job and this would affect cash flow."