Trust, privacy, and compliance

EU data residency and where your data lives

EU data residency is a core architectural principle of the Unless platform, not an add-on feature. Every piece of personal data from your customers or end users is stored and processed exclusively on cloud infrastructure within the European Union. This design directly supports GDPR compliance and the EU AI Act, and it removes the need for transfer impact assessments or reliance on adequacy decisions, because your data simply never leaves the EU.

Where your data is stored

All customer and end-user data sits on EU cloud regions. The core infrastructure runs on AWS in Ireland, with auxiliary workloads on EU regions of Microsoft Azure and Google Cloud. The platform stores different data types in separate, purpose-built systems, each with its own security measures:

The Privacy Vault, which holds the mapping between tokens and real identifiers, also stays in the EU. Only Unless’s own financial administration uses tools that may sit outside the EU, and those tools never touch end-user data.

How the Privacy Vault keeps data in the EU

The Privacy Vault provides twelve numbered privacy measures, including PII filtering, tokenization, data minimization, prompt anonymization, and EU data residency. When customer data enters the platform, it is filtered at ingestion and again at runtime, before any model call. Sensitive identifiers are removed or masked, and where personal context is required, the data is tokenized. The model only sees unreadable tokens, never the raw personal data.

De-tokenization happens inside the controlled Unless environment. The foundation model provider never receives raw personal data, which means the external model is not a sub-processor of your data. This design also solves a known GDPR problem: because the model never sees your data, there is nothing to delete from the model when a user requests erasure.

Sub-processors and compliance documentation

Platform sub-processors are EU entities of AWS, Microsoft Azure, and Google Cloud, all processing inside the EU. The full list lives in your Data Processing Addendum (DPA) and is updated when it changes. You can view the sub-processor list and data flows in the Compliance Center, along with the standard DPA, the security addendum, and other compliance documents. Your account manager can also send signed copies.

Unless does not train models on customer content, end-user conversations, or any other customer data. Living Knowledge improves through approved content and Team Feedback, never by feeding raw conversations into model training. Model providers are configured so that prompts and outputs are not reused for general training.

Retention and deletion

Data and logs are retained for a maximum of 365 days, or as long as needed to provide the service, support monitoring, and meet legal or contractual obligations, whichever is shortest. You can align retention settings with your own policies, and certain data types, such as conversation logs, can be limited or deleted according to your requirements.

When you need to process a data erasure request, open Trust > Privacy > Data subject requests and click New request. As the controller, you initiate the request. The Unless team verifies with a human in the loop, and the platform deletes everything related to the identifier across Living Memory, Conversations, and integrations. For data tokenized in the Privacy Vault, tokens orphan automatically when the underlying value changes or is deleted, so old references cannot be reused.

Conclusion

EU data residency is built into the Unless architecture at every layer, from storage to model routing. Your data stays in the EU, the model never sees raw personal data, and you retain full control over what the AI can access. This design lets you deploy AI agents without becoming personally liable for every model input and output, and it gives your DPO the documentation needed to sign off on a new pilot or an annual compliance review.