We develop and deploy AI features with a focus on safety, transparency and human oversight.
Principles
- Customer data is not used to train foundation models without explicit consent.
- AI outputs may be inaccurate — users should review before taking material action.
- High-impact decisions retain a human in the loop.
- We disclose when a user is interacting with an AI agent.
- Workspace AI actions require a separate user confirmation before selected content is sent to the configured provider. CSV insights use a sample of up to 20 rows, not the full dataset.
- Workspace AI access is limited to 10 requests per account per UTC day. A request uses an allowance even if the provider cannot complete it.
Do not submit personal, special-category or confidential information unless you are authorised to do so. AI tools are not enabled until the site owner configures a provider and applies the required database migration.
Your confirmation authorises processing to produce the requested response only; it is not consent for model training. External providers process submitted content under their own terms and configured controls, which must be reviewed before the customer pilot is enabled.
Prohibited uses
Do not use our AI features to generate content that is illegal, deceptive or harmful.
This document is a plain-language template maintained by NexGen eHub Limited. It is not legal advice. Please consult a qualified adviser for your specific circumstances.