From Human Expertise To Intelligent Agents: A Collaborative Framework For Industrial Design

The design of industrial products is a rigorous process that normally relies on multidisciplinary teams, manual analysis, and iterative refinement. As product complexity grows, traditional workflows struggle to deal with evolving requirements, regulatory constraints, and time-to-market demands. This paper presents the first step of a collaborative initiative between a company and an academic team, aiming to transform real-world engineering workflows into an intelligent, agent-based system powered by Large Language Models (LLMs). Leveraging platforms such as CrewAI, we structure design activities into modular, autonomous agents capable of understanding, extracting, and interpreting technical documentation. This work focuses on the transition from human-driven design to AI-assisted process. We analyze real documents, workflows, and specifications from the automotive antenna housing domain and show how these can be abstracted into agent compatible tasks. Each agent is designed to reflect the specific roles found in engineering teams, such as specification reading, requirement extraction, and technical interpretation. This approach lays the foundation for AI-driven automation in industrial design, showing how LLMs can support reasoning-intensive tasks, reduce cognitive load, and enhance collaboration. By bridging domain knowledge with intelligent systems, this work opens new pathways for scalable, flexible, and intelligent design processes across manufacturing sectors.