Which data of the customer tenant is used by AI/ LLM or Robin to support my AI Request?

Hello,
I’d like to clarify what specific information I need regarding my question: “What data from the client’s tenant is used by the AI/LLM or Robin to process my AI request?”
According to your myTOPdesk-Portal knowledge articles and the information provided by your consultant, can the AI only use data that has been created as knowledge articles within the tenant? Is that correct?  

In the first quarter of 2027, the AI will be optimized by TOPdesk, so that it can also read attachments such as Word, PDF, or Excel files from the client’s “Knowledge Items.”

In a third step, the AI will then be granted access to active and archived incident data.

Could you confirm your strategy regarding the AI’s use of tenant data?  

Please also explain how the AI handles imported emails or Microsoft documents that are marked as “sensitive” or “confidential.”