Abuse monitoring – data location, retention and human review for Switzerland North / West and EU Data Zone deployments

Hello,

We operate the following resources under our Microsoft Customer Agreement (accepted 18 October 2026):

• Azure OpenAI, resource "caai-resource", Switzerland North, Standard (regional) deployments
• Azure OpenAI, resource "caai-max", Data Zone Standard (EU)
• Azure OpenAI embeddings, resource "azure-ch-caai-resource", Switzerland North
• Azure AI Vision, Switzerland West

Our request for modified abuse monitoring was declined, so standard abuse monitoring applies. For our data protection documentation we need a written statement, per resource above:

1. In which country/region are prompts and completions stored for abuse monitoring, and for how long at most?
2. In which countries are the authorized Microsoft employees located who may review flagged content (for the Switzerland North/West resources in particular, since the public documentation only mentions EEA)?
3. Does Azure AI Vision (OCR / image analysis) store any customer content for abuse monitoring or human review? If yes, where and for how long?
4. Can you confirm that customer content (prompts, completions, images) is not used to train or improve Microsoft or OpenAI foundation models?
5. Which version of the Microsoft Products and Services Data Protection Addendum applies to our agreement?

Thank you