Sovereignty & Security

Designed for production: auditing, anonymization, governance.

Zaion ensures end-to-end confidentiality, traceability, and compliance, with sovereign hosting and encryption of sensitive data.

Compliance & Governance

AI Act

Compliance with the EU AI Regulation: system classification, model transparency, risk assessment, and comprehensive documentation.

GDPR

Data protection: anonymization of PII, data subject rights (access, rectification, erasure), record of processing activities.

DORA

Digital operational resilience: stress testing, business continuity planning, and management of risks related to third parties and cyber threats.

ISO 27001

Information security: security policy, access management, encryption, regular audits, and continuous improvement.

Data Localization & Protection

Data centers in Europe

Hosting on sovereign infrastructure to ensure data localization and compliance with European regulations.

End-to-end encryption

Encryption in transit and at rest for all sensitive data: conversations, transcripts, summaries, metadata, and logs.

IAM & Access Management

Granular identity and access management (IAM): multi-factor authentication, the principle of least privilege, action logging, and instant revocation.

Backup & Resilience

Sauvegardes automatiques multi-sites avec réplication asynchrone pour garantir la disponibilité et la récupération rapide en cas d’incident. RPO < 1 heure, RTO < 4 heures.

Guaranteed availability

0 %

Geographic replication

0 sites

Recovery Time Objective

< 0 h

Use of Data & Anonymization

Zaion adheres to a strict policy of minimizing and protecting personal data at every stage of the customer interaction lifecycle.

Uses by solution

  • AI agents: Conversational data used only for the current session; no long-term storage unless explicit consent is given for improvement purposes.
  • Agent Assist: real-time suggestions, anonymized logs for auditing and model improvement.
  • Conversational Insights: anonymized aggregation for statistical analysis; no personally identifiable information in the reports.

Anonymization of PII

Automatic detection and redaction of personally identifiable information (names, addresses, phone numbers, IBANs, contract numbers) prior to storage and processing by machine learning models.

Example of an anonymized transcript:

“Hello, my name is [***] and I live at [*******]. I would like to file a claim under my policy [*******].”

"Minimum Data" Policy

No unnecessary storage of inferences: model results are transmitted to consumer applications and deleted after use, unless there is a justified business need (audit, compliance, billing) with a limited retention period.

A safety issue?

Our experts are here to answer your questions.