Can the answer be automatically reformulated differently per customer

The same fact, different wording per portal

Yes, it can, and it happens continuously in practice. The underlying fact is established once in the supply chain file — for example the number of employees, the ISO 27001 status, or which of the ten measures from Article 21(2) of the NIS2 Directive have been completed. How that fact is worded in the answer field depends on how the customer has phrased the question. One buyer asks "Has an information security policy been established?", another asks "Do you comply with ISO/IEC 27001?" — both questions can be answered from the same underlying data, but the text that appears in the answer field is adapted to the wording and tone of that specific questionnaire.

This is not a matter of making the answer "look better" per customer. The system adapts the wording to the format — for example in full or abbreviated, with or without explanation, in the style of a yes/no question or of an open question — but the content remains bound to what is recorded in the file. With a question about the company's annual turnover it makes a difference for example whether a portal asks for an amount in whole euros or for a bandwidth; the underlying turnover does not change, only the presentation. The same applies to a question about in which countries the enterprise has establishments: some portals want a list, others a number.

Where the line is drawn: confirmation remains necessary

Every proposed answer — including a different wording of a known fact — is confirmed yes or no per question before it is submitted. This follows from the platform's method of operation: submission is never automatic without that confirmation, and no answer is ever proposed for which the file has no underlying fact. With a question that touches on the basic module of the EFRAG VSME standard, for example, the wording is adapted to the format of the portal, but the figure or measure itself always comes from the file, not from the customer's question text.