With language assistants, automatic translators and new record-breaking language models appearing every week, it seems like anything is possible, doesn’t it?
We have been using Natural Language Processing (NLP) techniques for many years in more than 60 projects in the automotive and insurance sectors for the quality assurance of software. Examples are the automatic checking of requirements, test generation from user stories, and automated traceability analyses.
This results in a somewhat more differentiated picture.
In this presentation we will show what is state of the art, what is not yet possible in practice, and what (in our opinion) will never work. With this we want to show the possibilities of text analysis without falling into unrealistic expectations or even buzzwords.