Agreed that they're not black-boxes and there's a lot of assumptions you can carry over, though. The NCAP rating of the donor vehicle won't be valid any more, but all that tells you is how well its siblings survived destructive testing - it's likely to perform similarly.
TLDR: it's not quite that simple, but assuming you're not driving like an idiot it'll probably be fine.
Cars are designed to operate in a certain range of weights, as they get loaded with drivers, passengers, cargo, etc.... so long as the distribution is roughly the same, then the careful tuning of the suspension and handling, etc... should all still mostly work the same.
Sure, it might be a bit more sluggish if it is significantly heavier, but that should be about it. You might b able to largely solve that problem by just putting some stronger springs in the suspension, which might already exist in the third party market for other people who are doing other kinds of work on the same type of vehicle.
Of course, you’ve got to actually do the work to make it right, but keeping the same basic weight distribution will probably solve at least 80% of the problem.
With regards to components being really heavy, consider that the engine and the transmission are two of the heaviest components in most any car. If you can eliminate those in your BEV conversion, that gives you quite a bit of weight to work with for the new heavy components.
Furthermore, in the case of an EV conversion you can basically just make sure the EV power-train does not occupy physical space that the ICE power-train did not and your crumple zones and airbags will work fine.
Edit:
Making batteries that don't catch on fire is a solved problem. Getting those electrons to the motor without starting a fire is a solved problem. You really only get fires when you're trying to push the performance envelope which is generally a low priority for commercial conversions.
Cars are not hyper-complicated non-deterministic black boxes. It is possible to modify them without causing unforeseen consequences. Just because you may personally lack the expertise to understand how the systems interact does not mean that nobody can understand it. A qualified engineer with an automotive background should have no problem creating an EV conversion that is every bit as safe as the base vehicle though it will probably be slightly less efficient in terms of materials and labor usage than a conversion for the same vehicle designed by an OEM team for manufacture at scale. The OEMs don't have people that are magically smarter, or different laws of physics if anything they will have a harder time engineering an equivalently performing system because they have many more constraints to optimize for than a conversion company.