Zendesk

Zendesk says third-party LLM features are not trained on Zendesk customer data.

Zendesk says third-party AI processing uses zero data retention beyond the ephemeral time required to return a response.

Customers can decide whether their data may be used for product-improvement purposes.

Zendesk embeds privacy-notice tooling for messaging customers instead of leaving disclosure entirely out of product setup.

Zendesk proprietary models may be trained on customer data to the extent instructed by the customer.

Zendesk AI still processes tickets, transcripts, and customer-support context when those features are enabled.

Zendesk AI still relies on proprietary models, vetted LLM providers, integrations, and ordinary service providers.

Ordinary support-platform sharing with vendors, legal recipients, and customer-configured integrations still applies.

Zendesk says service data processed by Zendesk AI remains subject to existing security and privacy commitments.

Zendesk performs security reviews on vendors with access to systems or service data.

Data locality choices and least-privilege access controls help customers constrain who sees service data.

Redaction and deletion controls let customers limit what survives or gets exposed to downstream tools.

Marketing, analytics, and website tracking still apply outside ticket-level controls.

Agent, end-user, and support activity still generate telemetry and service-usage metadata.

Zendesk offers embedded privacy notices for messaging deployments.

Standard redaction and privacy-enhancing controls reduce some downstream exposure even after collection.

Zendesk is less advertising-centric than consumer social platforms, and many higher-risk settings are customer-administered.

Advanced retention policies and some stronger controls may depend on add-ons or plan eligibility.

Backups, compliance, and account-specific setup still affect how long some support data persists.

Zendesk offers ticket and end-user data deletion schedules out of the box for all customers.

Customers can permanently redact sensitive ticket content using standard tooling.

Third-party AI processing is described as zero retention beyond the ephemeral response window.

Admins can configure AI features, redaction, and deletion schedules.

Zendesk offers multiple authentication, access, and least-privilege controls for agents and admins.

Customers can choose data-locality options and message-level privacy notice behavior.

Customers can decide whether data may be used for product improvement and can redact or delete sensitive content.

Some stronger controls still depend on eligible plans or add-ons.

Customer instructions and configuration still determine how much data actually reaches AI features and integrations.