Topic

AI Agent Empowers Manufacturing and Educational Publishing: A Scenario Revolution from Intelligent Design Assistance to Knowledge Services

As the CPO and CTO of Suoyun AI, I will share two industry scenarios that I personally led. 1、Manufacturing Scenario: Intelligent Material Selection AI case, in-depth analysis of how AI Agent breaks through the traditional manufacturing data silos and efficiency bottlenecks to realize the intelligent upgrade of the whole chain: For an electronic manufacturing enterprise, “low efficiency of material data entry, difficult to match the design requirements, and reliance on manual substitution decision-making” pain points, we based on self-developed AWE engine and large model technology to build a new AI agent that covers material analysis, demand transformation, comprehensive query, comparison and substitution analysis. Based on the self-developed AWE engine and big model technology, we build an AI Agent matrix covering material analysis, demand transformation, comprehensive query, comparison and alternative analysis: Intelligent analysis and library building: Through the “material information collection and entry Agent”, it supports automatic analysis of multi-language Datasheets (PDF/text), and the AI Agent can be used to analyze and analyze material data. Classification The accuracy rate of AI classification reaches 98%, the efficiency of attribute extraction is improved by 5 times, and automatic synchronization with PDM system data is realized, with 70% less manual intervention. Precise transformation of requirements: Based on “Material Design Requirements Understanding Agent”, engineers' fuzzy requirements (e.g. “high-temperature resistant small package resistor”) are transformed into structured parameters, semantically aligned with Chinese and English technical documents, and the efficiency of requirements clarification is improved by 60%, and the accuracy of technical parameters matching reaches 95%. The efficiency of requirement clarification is increased by 60%, and the matching accuracy of technical parameters reaches 95%. Optimization of substitution decision-making: Through “Material Comparison and Substitution Analysis Agent”, linking ERP and WMS system data, combining cost, inventory, performance and other multi-dimensional weighting scores, generating executable alternatives, shortening the substitution decision-making cycle from 3 days to 1 hour, and reducing the procurement cost by 15%. The product deeply integrates manufacturing know-how, builds the technology closed loop of “AI+process automation+multi-system synergy”, and becomes the benchmark practice of digital intelligent transformation of manufacturing industry. 2、Educational Publishing Scenario: AI Thesis Full Scene Intelligent Service creates an AI service matrix for academic journals for X Publishing House, and increases the reading efficiency of scientific research thesis by 3 times and user satisfaction by 40% through the technologies of AI academic search, AI thesis translation, thesis speed reading (in Chinese and English), multi-document mind mapping, thesis popularization video generation, thesis abstract video generation, etc. The project deeply integrates publishing industry standards and becomes a standard practice of digital intellectualization transformation in manufacturing industry. The project has become a technology and commercialization benchmark in the field of “AI+Publishing” by deeply integrating the publishing industry standards. This speech will focus on the technical architecture design (such as Agentic Workflow engine), cross-industry methodology (scene binding and cognitive value closed loop), as well as the subscription business model for large-scale replication, providing a one-stop reference for global enterprises from the AI technology landing to the validation of commercial value.