Where are documents processed by AI?
AI processing at amaise is strictly regional — data does not leave the assigned geographic zone:
Environment | AWS region | Azure OpenAI | Other services |
EU | Frankfurt | EU data zone (models exclusively in the EU) | OCR in the EU (Google Document AI) |
US | Ohio | US data zone (models exclusively in the United States) | OCR in the US (Google Document AI) |
CH | Zurich | Switzerland North (models exclusively in Switzerland) | OCR in Switzerland (Azure AI Document Intelligence) |
Important guarantees:
No cross-region data transfer between environments
Text extraction (OCR) runs in-region: Swiss documents are read in Switzerland (Azure AI Document Intelligence), EU and US documents in their own region (Google Document AI)
Azure OpenAI deployment SKUs ensure data processing remains within configured regions
The OCR services keep documents only temporarily for processing (Google Document AI: at the latest two days; Azure AI Document Intelligence: at the latest 24 hours); under Microsoft's terms, Azure OpenAI may retain prompt and completion content for up to 30 days for abuse monitoring, never for training
Customer data is not used for model training by default
The optional legal research sources (for example web research via Perplexity in the USA) are only used when activated for a workspace and receive search queries only — no documents. LegalPilot, when activated, sends the user's question to Perplexity