In several domains, there is a growing demand for automation software for enhancing process efficiency and reliability. The railway industry is a notable example, given its safety-critical nature and the importance placed on reliability and data integrity. To tackle the complexity of modern trains, engineers must use different notations to specify the intended system and subsystems during development. Managing these different notations and ensuring accurate translations between them poses significant challenges. Manual revisions and translations are time-consuming, costly, and prone to human error, potentially introducing faults into the system. Consequently, automating the extraction of relevant information from these documents can help address these challenges, leading to improved efficiency and accuracy in the development process.In this thesis, we design and developed a NLP-based framework for the semi-automatic translation of semi-structured texts into structured data.The framework focuses on ensuring integrity and reliability of the translated data.We validate the proposed framework on an industrial use case from the railway domain provided by our partner Alstom Rail Sweden AB.