Large Language Models Assistance in core Model-Driven Engineering activities

Abstract

Recent research explores the use of Large Language Models (LLMs) and conversational agents to enable automation and assist modeling tasks across the Model-Driven Engineering (MDE) lifecycle, including design, model management, and evolution. This paper presents a Systematic Literature Review of 96 peer-reviewed studies on the use of LLMs in MDE, aiming to consolidate existing knowledge, highlight emerging patterns, and identify key limitations and research opportunities for future advancement. We systematically analyze the literature to identify which core activities are currently assisted and the types of modeling languages and artifacts involved. We further examine how LLMs are used to support these activities and the role of human involvement in the loop. Finally, we address the maturity of these solutions in terms of technology readiness level, domain applicability, and the availability of supporting tools.