Solutions
Speech Technology
Products
Resources
Company
Contact

VoiceInteraction AI Super Cluster

Data Processing and Security Information

Data Processing and Security Information

Last update: Oct 2026

Last update: Oct 2026

Document status: This customer-facing template requires approval by VoiceInteraction legal, privacy, and information-security teams before distribution. It provides operational information and is not legal advice or a claim of certification against a particular standard.

Document status: This customer-facing template requires approval by VoiceInteraction legal, privacy, and information-security teams before distribution. It provides operational information and is not legal advice or a claim of certification against a particular standard.

ON THIS PAGE

ON THIS PAGE

Data Processing and
Security Information

Service

VoiceInteraction Central Knowledge Manager is part of the VoiceInteraction AI Super Cluster, administered by VoiceInteraction IT and hosted in a high-security, controlled-access data-centre environment.

Data processing

Input content and generated results are processed transiently to perform the requested operation.

  • CENTRAL is designed not to persist transcripts, input text, prompts, generated results, authorization values, or provider response bodies in its Elasticsearch operational records.

  • Results may remain temporarily in process memory until retrieved or expired.

  • The local AudimusServer installation may save the resulting title, summary, or other requested metadata as part of normal product operation.

  • Where an external LLM provider is used, its processing and retention are governed by the applicable provider configuration, contract, and data-processing terms.

Operational records

CENTRAL stores content-free operational information required for authentication, security, quota enforcement, support, service operation, and usage accountability. This may include:

  • tenant, instance, credential, request, and job identifiers;

  • operation type and language;

  • provider and model;

  • status and bounded error category;

  • character and token counts;

  • estimated cost and latency;

  • provider request identifier and timestamps.

Operational records are retained according to the applicable service configuration, contract, and legal requirements. Authorized access, correction, deletion, restriction, or export requests must follow the agreed support/privacy process.

Security

Communication uses HTTPS with hostname and certificate validation.

  • Every installation authenticates using a tenant-specific credential.

  • CENTRAL retains a verification hash, not a recoverable copy of the tenant credential.

  • After local configuration, the credential is stored server-side by the local Knowledge Manager and is not returned by configuration APIs to the browser.

  • Provider API keys remain exclusively in CENTRAL and are not distributed to customer installations.

  • Tenant and route identities must match.

  • CENTRAL administration is restricted to the service host and authorized operators.

  • Automatic fallback to a local provider is disabled by default.

  • Suspected credential exposure requires immediate revocation, rotation, and notification through the security contact.

These measures do not constitute a guarantee of absolute security or availability. Any certification, compliance, data-locality, or provider non-retention commitment must be established separately in the applicable contract and supporting assessment.

Customer responsibility

Only submit content that may lawfully be processed for the requested purpose. Avoid unnecessary sensitive or prohibited content. Applicable controller/processor roles, provider subprocessors, processing regions, retention commitments, and international-transfer mechanisms are defined by the governing contract and data-processing agreement.

When requesting support, provide only content-free identifiers such as tenant ID, credential ID, instance ID, request ID, job ID, timestamp, and error code. Do not include original content, credentials, authorization headers, prompts, or generated results unless an approved secure process explicitly requires them.

Questions about Data Processing and Security Information?

Questions about Data Processing and Security Information?

If you have questions about this Data Processing and Security Information, contact VoiceInteraction through the appropriate contact channel.

View all contact options →

ON THIS PAGE

Privacy Policy & Cookies