Thank you.
Thank you for the question.
The software's corpus is supplied with phrases and terminology from the government. These are drawn from current government websites. The corpus was developed by the National Research Council of Canada.
In terms of assessing content, I believe you want to find out what kind of results we obtain once the data has been entered. It's really all about looking at the statistical recurrence of data. If a match in content, for example "dog means le chien" comes up quite frequently, that is the result that's going to appear. The content has to be re-worked and validated by professional language experts, who will also enure quality control. After that, the corpus is sent to the National Research Council, which does what's known as retraining of the engine to search for those reoccurrences. We can add the content we want to the corpus so as to obtain the best results possible.