You have a bit of a chicken and the egg situation here. I would say that it is absolutely worth doing CS and maths and CS honours, etc. if you are actually going to use it. Of course, you are far more likely to use it if you actually have the qualification.
The only real way out of the dilemma is to really figure out what you want to be doing for your career - this is non-trivial since until you will only have a better idea once you do the work.
Being indispensable isn't really a good argument - as Sinbad said, much of what makes you indispensable will come out of how you do your job.
Some high level issues:
- Where do you want to work? If you want to stay in SA, there are far fewer opportunities that need a CS/computing degree. If you really want to go to Silicon Valley, or such, getting a CS background is worth it (personally I think it is still "worth it" for most people generally, but probably not if you already have an honours/masters informatics degree).
- What kind of work do you find interesting? Algorithms, maths, CS theory, compilers, ML? If you really know that you want to work at this kind of technical level, then perhaps going the CS/computing path is best.
- Do you see yourself as a career developer? If not, I think what you have will get you to the business side or management side of many businesses, so CS/computing isn't worth it. If you want to start your own business, will it be generic software development, or do you want to build something around new technology you've developed? The latter would benefit from a CS/computing degree, but would also require that you are exceptionally talented in this area to really compete.
I do agree with what Messugga said about having 2x degrees rather than a masters and/or PhD. From the perspective of an outsider reading your resume it may look as though you didn't have focus, or weren't capable of taking a subject to the masters level (I know one or two career students who have 3-4 degrees, but no masters for this exact reason - 4 degrees and not particularly bright).
I do shudder to think of what an ML course that doesn't need maths looks like. Regardless of whether it is a requirement for a particular module, if you want to something ML or data science related at the postgrad level or for your career, maths and stats is highly recommended.