Whether you're a developer, consultant or a customer, it's always difficult to go through the entire documentation to find a specific thing.
The solution is to have a ML model that can answer questions on a given passage/corpus. There are state of the art ML models that are trained using the relationship between the words as the features to answer questions in a given context. Hence it doesn't matter if there are non-english words in the corpus.
The model can answer any type of questions posed, including: what?, how?, which?, ...
Having such a solution will help one find the relevant data faster and provide a better user experience.