Towards LLM Agents for Model-Based Engineering: A Case in Transformation Selection

In Model-Based Engineering, practitioners frequently face the challenge of selecting appropriate tools from a large number of options. LLM-based agents depend on Large Language Models to autonomously select and apply software tools to perform specific tasks. Although LLMs have already been applied to support various MBE activities, considering LLM-based agents to autonomously assist users of MBE tools remains underexplored — particularly in industrial MBE environments where only medium-sized on-premise LLMs can be used due to security or data-privacy policies. We investigate the potential of LLM-based agents for MBE, starting with model-to-model transformation as a core MBE technique. We propose an approach based on complementary mechanisms: (i) a model transformation server and an LLM agent with dedicated tools for each available transformation; (ii) a tool retrieval technique based on a tool relevance score computed by an LLM. Our comparative study on a contributed transformation dataset shows that the newly proposed LLM agent responds more accurately to user instructions.