Jian Chen

Faced with the growing scale of multilingual content, compressed delivery timelines, and rising quality requirements, traditional TMS and siloed AI translation solutions can no longer meet the demands of enterprise-level localization production. This session will explore the architecture positioning, core capabilities, and practical implementation of next-generation TMS, discussing how AI agents can be deeply embedded into the translation workflow to perform task understanding, automatic planning, multi-agent collaboration, terminology and memory reuse, style control, format preservation, automated quality assurance, and result delivery. Drawing on the practice of YiCAT Agent, we will analyze how AI has evolved from a standalone translation tool into a controllable, traceable, recoverable, and deliverable intelligent production node. This provides a reference for enterprises looking to build an intelligent multilingual production system that harmoniously integrates people, language assets, AI, and processes.