Trust
Privacy
Version 1.0 · Last updated 2026-05-27
GDPR is non-negotiable in Europe. The EU AI Act adds another layer on top, especially for customer-facing AI. Unless lets you deploy an agent without becoming personally liable for every model input and output.
Key concepts
Privacy Vault
Twelve numbered privacy measures, including PII filtering, tokenization, data minimization, prompt anonymization, and EU data residency. Sensitive identifiers never reach the model. De-tokenization happens inside the controlled Unless environment; raw personal data never reaches the foundation model provider.
Filtering at ingestion and runtime
Personal data is filtered when content enters the platform and again at runtime, before any model call. Identifiers are removed or masked. Where personal context is required, it is tokenized through the Privacy Vault and the model only sees tokens.
No EU data transfers
Personal data on the platform stays inside the EU/EEA by design. Unless does not rely on adequacy decisions or transfer impact assessments for platform data, because the data never leaves the EU.
Sub-processors
Platform sub-processors are EU entities of AWS, Microsoft Azure, and Google Cloud, all processing inside the EU. The full list lives in your DPA. Business tools used for Unless’s own administration may sit outside the EU; those tools never touch end-user data.
No training on customer data
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.
PII filtering
Names, emails, phone numbers, and other personal data are detected and removed by default before they reach the model. Whitelist exceptions for fields you need the model to see.
Tokenization
Sensitive identifiers are replaced with tokens. The agent works with tokens. Only the Privacy Vault holds the mapping back to the real identifier.
Consent management
Configure whether tracking and personalization require explicit consent. Two modes: consent not required, or consent required (with sub-options for cookieless experiences).
Data subject requests
Built-in workflows for data access, correction, and erasure requests from end users. Track each request from submission to fulfillment.
Data residency
Personal data and the Vault stay in the EU. Sub-processors are listed in your account and updated when they change.
What you can do here
- Configure PII filtering and the whitelist
- Switch consent mode and sub-options
- Process data subject requests
- View sub-processor list and data flows
- Export DPA and security documentation for your DPO
When to use it
- During a new pilot when your DPO needs to sign off
- When you receive a data subject request from an end user
- When you change which fields the agent sees from your CRM
- During an annual compliance review
How it works
When customer data enters the platform, the Privacy Vault filters and tokenizes before anything reaches the model. The model sees scrubbed input and generates output. The Vault re-attaches identifiers only when the response needs them (e.g. to address the customer by name).
Frequently asked questions
How do I process a data erasure request?
Open "Trust > Privacy > Data subject requests" and click "New request". As the controller, you initiate. The Unless team verifies with a human in the loop. 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.
How do I add a field to the PII whitelist?
Open the PII whitelist and add the field. The agent will see it in plain text. Only add fields the model actually needs.
How do I switch on consent-required mode?
Open the Consent management settings and switch to "Require consent". Pick whether cookieless experiences should still show.
Where is my data stored?
All personal data on the platform is in the EU. Core infrastructure runs on AWS in Ireland, with auxiliary workloads in EU regions of Microsoft Azure and Google Cloud. Only our own financial administration uses tools that may sit outside the EU, and those contain no end-user data.
Does Unless train AI on my data?
No. Customer content, end-user conversations, and any other customer data are never used to train models. Living Knowledge improves through approved content and Team Feedback, not through model training.
Can I get a DPA?
Yes. The standard DPA, the sub-processor list, the security addendum, and the rest of the underlying compliance documents live in the Compliance Center. Your account manager can also send signed copies.
What is a cookieless experience?
An experience that does not require tracking. Can show before consent for visitors who have not opted in yet. Useful for compliance with consent-required setups.
