The introduction of the European Commission’s ‘Safe and Sustainable-by-Design’ (SSbD) framework represents a catalyst in the wide adoption of SSbD practices and it has spurred an increasing interest in displaying practical implementations among policymakers, academia, and industry players, such as the chemical sector. Conducting such a type of assessment at early stages of the design process is far from trivial due to the lack of data or the need of lengthy modelling pipelines. This study proposes a novel modelling workflow that combines the power of Large Language Models and machine learning methods to bootstrap the data gap filling in the early stage of development of new chemicals by automating the generation of Life Cycle Inventories. The workflow is meant to produce ready-to-use proxy data that can be later used by any LCA computing engine to be enhanced or analysed in detail.