This study examines the paradigm shift in Chinese web literature intellectual property (IP) production, driven by rapid advances in Artificial Intelligence Generated Content (AIGC). Grounded in Henry Jenkins' "Convergence Culture" theory and Actor-Network Theory (ANT), the research conceptualizes AIGC not merely as a passive tool, but as a pivotal "actant" that propels web literature from a "human-interaction" model (Convergence 2.0) to a "human-machine collaborative system" (Convergence 3.0). This study asks: How does AIGC reshape the narrative production and power structures of web literature IPs? Using evolutionary case analysis, this study traces the transformation across three stages by analyzing the practices of leading platforms. It constructs and validates a dynamic narrative workflow model—trigger–generation–screening–feedback— Stage 1, represented by Tomato Novel, illustrates a data-driven workflow where behavioral metrics construct a "database authority, " gradually transforming authors into "data product managers" through algorithmic governance. Stage 2, exemplified by Yuewen's "Miaobi" model, introduces AI-generated creation, enabling generative exploration while risking the dilution of artistic individuality and the erosion of textual “aura” through algorithmic standardization. Stage 3, conceptualized through Zhihu Salted Fish, envisions a fused data-AI intelligent workflow characterized by a "diagnosis-generation closed loop, " where extreme algorithmic centralization paradoxically coexists with maximal democratization of user participation. The core findings reveal that this evolutionary shift establishes a radically new mechanism of "narrative management, " marking a fundamental departure from traditional static "work management." This transition manifests across three dimensions: the management object shifts from a closed text to an open, expandable "narrative system" (comprising narrative databases, generation engines, and multimodal interfaces); the management process transforms from a linear assembly line to a dynamic, real-time collaborative network; and the management goal migrates from delivering a finite product to sustainably operating an IP narrative ecology and scaling personalized experiences. Consequently, this technological and structural upheaval necessitates a historic reconstruction of human labor within the publishing industry. The traditional web novel editor must transition from a passive "gatekeeper of knowledge" into a more proactive "narrative engineer." To operate effectively in this AIGC-mediated environment, narrative engineers must master four core competencies: "architectural competency" to design self-consistent, expansive narrative universes; "prompting competency" to translate abstract aesthetic demands into precise machine-executable instructions; "data competency" to interpret algorithmic metrics and optimize narrative iterations; and "integration competency" to seamlessly stitch AI-generated fragments with human creativity, ensuring final artistic coherence. Ultimately, this study contributes a meso-theoretical framework bridging convergence theory and digital publishing practice. It offers strategic pathways for publishing institutions—shifting from product delivery to ecological operation, restructuring organizational workflows, and cultivating interdisciplinary "narrative management" talents—while highlighting future research directions in narrative ethics and cognitive mechanisms of human-machine co-creation.