In the midst of today's digital revolution, businesses face a common challenge: incorporating standards, often in PDF form, into their digital ecosystems. The push to digitize product development and procurement processes has underscored the need for efficiently converting traditional PDF documents into usable digital formats. Often originally authored in Word documents, standards are subject to diverse interpretations. While new authoring guidelines may enhance machine readability for new standards, the primary focus lies in digitizing the thousands of existing legacy standards. In this presentation, SAE International will share our lessons learned around digitizing standards using natural language processing models. We provide guidance about what to look for in standards structure to determine whether a standard is a good fit for natural language processing models or not. We also share lessons learned about upfront planning that is critical to ensure that the digital artifacts created will support your use cases.
