Review of AI agents in engineering design workflows

Technology related to artificial intelligence (AI), especially agentic systems, has attracted significant interest and rapid advancement in recent years. With this growth comes an increasing potential for integrating agents into engineering design workflows. This paper examines the current state of AI agent technology from an engineering design perspective. Descriptive bibliometric analysis characterizes recent research trends, showing rapid growth in engineering-related research despite comparatively limited publication volume. A consistent boundary for agentic behavior is established to distinguish task-specific AI and near-agentic workflow automation from systems demonstrating more direct agentic participation. Recent single- and multi-agent workflows demonstrate meaningful capabilities, although their implementation in engineering design remains limited. Contributing technical, ethical, and human-centered challenges are discussed alongside their implications for broader adoption. The technology's broader impacts on engineering are then considered, with an emphasis on the potential for agents to assume more critical roles in design workflows as these barriers are addressed.