Integrating definiteness and the whole indefinite pronoun system with its various functions into noun morphology instead of having things like case could be an interesting approach to noun morphology - it seems to me the order by which conlangers go for noun morphology beyond the derivative morphology is a hierarchy something like
number (maybe fused with gender)
possession marking (head or dependent marking)
other cases (maybe fused with gender)
possessive affixes
definiteness marking
noun class / gender (maybe fused with number)
This is not particularly bad or anything, but we could do something else with nouns than that. Some Native American languages offer us the idea of marking for obviativeness/proximativeness, which interacts with the verb and the more general discourse in interesting ways. Few conlangers make Native American languages, however.
The last in the hierarchy above is gender, which any Bantu-style language would almost necessarily be present.
Now, I've often gone and linked the typological classification of indefinite pronouns that Apollo Hogan wrote way back. To this, we could add some definite pronouns and determiners - demonstratives, maybe articles (if we go so far as to distinguish 'that', 'this' and 'the'; seems 'the' may easily turn superfluous). From this point on, 'the/a/any/some/...' represents whatever system you come up with from that classification.
Incorporating that whole system of 'the/a/any/some/...' into the noun, possibly in combination with possessive affixes (either giving {the, a, any, some, no, every, ...} * {my, your, his/hers, our, ...} or {the, a, any, some, no, every, ...} + {my, your, his/hers, our, ...} could give interesting results. Let's further permit a "light recursion", having the third person possessive suffixes further be marked for a less granular set of distinctions - merge a few of the different 'anies' and 'somes' that you have for that suffix, and maybe forbid certain combinations ('than any X of than any X' seems unlikely to ever be needed, i.e. forbid double indefinite standard of comparison. I find it likely that you'll ever need the possessor to be the indefinite standard of comparison, but you could of course permit having the marking go there nevertheless for whatever reason, or heck, require doubly marking it.)
Since indefinite pronouns often have somewhat overlapping functions (just check the amount of overlap in the example systems section of the link), this gives us a somewhat more overlapping system than the typical case system - which is a nice effect, in my opinion.
Even more interest could maybe be created by having different noun classes divide up this functional space in slightly different ways, maybe even having different numbers of divisions of it.
