Can the system handle a questionnaire in two languages

The language of the question is not the issue, recognition is

The system recognizes questions based on content, not on language. A question about "Onder welk registratienummer staat de onderneming ingeschreven" is recognized whether it appears in Dutch or in English as "company registration number" — the system links the question to the underlying data in the supply chain file, not to the literal wording. Mixed questionnaires, where one section is in English and another in Dutch, are common with international customers who partially localize their portal and partially do not. In practice, this works just as well as a list in a single language.

What does matter is the question behind the question. "Number of employees" and "Aantal medewerkers" are the same question in two languages, and lead to the same answer from how many employees the undertaking has. But "Data protection officer" and "Functionaris gegevensbescherming" resemble each other and are that as well, whereas "Business continuity" and "Bedrijfscontinuïteit" do touch on the same subject as the continuity plan, but can ask for a slightly different level of detail per portal. Language is therefore rarely the issue; the precision of the underlying question is. When in doubt about what a specific English term precisely means, that is something to confirm per answer — the system proposes, the user approves or rejects.

Recognition works at question level, not at document level

This does not follow from a single legal rule but from how recognition of questionnaires works: each question is analyzed separately and linked to a category in the supply chain file, such as the data from the Trade Register Act 2007 or the ten measures of article 21(2) of the NIS2 Directive. Because that linking is based on meaning and not on exact text, the language of the original question makes no difference to whether an answer is found.