What I propose in the present article are some theoretical adjustments for a more coherent answer to the legal "status question" of artificial intelligence (AI) systems. I arrive at those by using the new "bundle theory" of legal personhood, together with its accompanying conceptual and methodological apparatus as a lens through which to look at a recent such answer inspired from German civil law and named or partial legal capacity. I argue that partial legal capacity is a possible solution to the status question only if we understand legal personhood according to this new theory. Conversely, I argue that if indeed lends itself to being applied to AI systems, then such flexibility further confirms the bundle theory paradigm shift. I then go on to further analyze and exploit the particularities of to inform a reflection on the appropriate conceptual shape of legal personhood and suggest a slightly different answer from the bundle theory framework in what I term a "gradient theory" of legal personhood.