Abstract: Artificial Intelligence (AI) has created significant opportunities for technological innovation in enterprises. From the perspective of knowledge flow theory, this study conducts a comparative case analysis of two AI startups from different industrial backgrounds to explore the mechanisms of AI-driven innovation processes. Research findings: (1) The AI-driven innovation process follows the path of “knowledge production—AI-driven market opportunity insight—knowledge acquisition—knowledge creation—knowledge application”,wherein the knowledge creation stage achieves the chimeric recombination of AI and domain-specific knowledge through the path of “training model—model prediction—experimental verification—iterative optimization”; (2) The innovation process is influenced by both internal and external factors: internal factors include founders’ knowledge endowment, compound talents, and organizational structure, while external factors encompass partnership networks and innovation environment factors such as technology, policy, and market conditions; (3) Enterprises with different industrial backgrounds follow similar knowledge chimeric paths but develop differentiated industry-specific models, while enterprises with different entrepreneurial backgrounds adopt differentiated knowledge acquisition strategies based on their characteristics.This study is helpful to deepen the understanding of the mechanism of AI-driven innovation process and provide practical enlightenment for enterprises to accelerate AI-driven innovation.