Manufacturing quality audits are critical to ensuring high product standards in mass production environments. However, traditional audit processes are labor-intensive and rely heavily on human expertise, posing challenges to maintaining transparency, accountability, and continuous improvement in complex global supply chains. To address these challenges, we propose an intelligent auditing system based on large language models (LLM). Our approach introduces three key innovations: a dynamic risk assessment model that streamlines audit procedures and optimizes resource allocation; a manufacturing compliance co-pilot that improves data processing, retrieval and evaluation for a self-evolving manufacturing knowledge base; and a Re-act framework common element analytics agent that provides real-time custom analytics to empower engineers with insights to improve suppliers. These improvements significantly increase audit efficiency and effectiveness, with test scenarios demonstrating an improvement of more than 24%.
