Organizations increasingly adopt Artificial Intelligence (AI) altering employees’ solution search behavior. In fact, AI may exacerbate or mitigate key search outcomes such as distance and effort according to prior work. We reconcile prior theoretical propositions and extend them by theorizing about the nature of the search agent (AI as a substitute vs. complement) and the underlying task setting (well- vs. ill-structured). We categorize four types of AI with different degrees of bounded rationality required to search that decide upon the effect of AI for search outcomes. To account for the interdependencies of these four types across time, we then adopt a process perspective. In doing so, we reveal how splitting search processes between different human-AI combinations helps to maximize the potential of AI for individuals’ search behavior. Overall, we contribute to the Carnegie School concept of search by extending its assumptions to emerging non-human, ‘intelligent’ agents such as AI.