This paper presents research on how human-algorithm interactions facilitate tacit knowledge conversion in knowledge-intensive industries. The study is based on a multi-case analysis of three Chinese patent medicine (CPM) manufacturing firms, the research investigates how algorithms, interacted with human expertise, contribute to the conversion of tacit knowledge in the CPM sector—encompassing sensory skills, diagnostic intuition, and learned heuristics—into structured, usable knowledge for algorithmic systems. Our findings highlight the iterative and co-constitutive nature of human-algorithm interaction in tacit knowledge conversion, in which both human expertise and algorithmic capabilities dynamically inform each other.