"It is not an issue of writting the most efficient and shortest code possible,"

Sometimes - just sometimes - this really is the Big Issue.
In these cases, you will need a highly-skilled Assembler programmer who can get you the absolute maximum performance from the absolute miniumum code.

But note that the key issue here is having a highly-skilled programmer using Assembler - an idiot using assembler will still turn out rubbish; a highly-skilled 'C' programmer may well be able to produce tighter, faster code than a stupid Assembler programmer.
As I've said time and again before, Assembler is not magic; simply switching to assembler will not automatically make your code tighter & faster - that's all down to your skill in using the language.

"it is an issue of getting your job done and to maintain it for future updates and bug fixes easily."

Yes, this is a very important issue, and one which must be considered very carefully in deciding the 'C'/Assembler partitioning of a Project.
As Erik (I think) has already said, it depends on whether the up-front Engineering cost, or the ongoing component cost is the most significant.