How AI features use your data
- AI features answer from your tenant’s data and cite their sources.
- Requests go through the CloudVision AI gateway, which applies your tenant’s AI policy, masks secrets, enforces quotas and records an audit entry.
- In our hosted service, your Customer Data is never sent to an external AI provider. AI features use only self-hosted models; until one is available they are off and the products give their standard answers.
- We do not use your Customer Data to train models.
- Prompt and answer text is not stored in audit logs unless you turn that on; a hash of the prompt and token counts are stored.
Your controls
We can turn AI off for your tenant and set budgets and quotas per tenant, user and feature on request. In a private deployment you control these settings yourself, including the model provider and the highest data classification that may reach it.
Your responsibilities
- Review AI output before acting on it. AI can be wrong.
- Don’t paste secrets or personal data into AI prompts unless your policy allows it.
- Approvals of changes are your decision. AI can recommend changes but cannot approve them.
- Comply with any laws that apply to your use of AI, including transparency obligations to your own users.