Draft: this page is waiting for legal review and may change.
Last updated 20 October 2026
Acceptable Use
1. Honesty about the AI
The AI must tell customers it is an assistant; it may never be set up to claim to be a person.
2. No advice outside a business's knowledge
The AI answers only from a business's approved knowledge. It must not be set up to give medical, legal or financial advice the business hasn't explicitly approved.
3. No misuse of channels
Businesses may not use Ariium to send spam, unsolicited marketing outside a channel's rules, or anything that breaks the policies of WhatsApp, Meta or any other connected platform.
4. No getting around identity checks
Businesses may not switch off identity checks for sensitive topics, or set up the AI to share data it hasn't confirmed belongs to the person asking.
5. Enforcement
We may suspend a workspace that breaks this policy, with notice where practical, to protect other businesses and their customers.