Trust
No back door, by construction.
The strongest statement on this site about the AI, and it is structural rather than a policy: the assistant queries as the asker, through the same permission checks.
One permission model
A second one would eventually diverge from the first, and the divergence would be a breach. So there is only one.
Customers are scoped to themselves
The assistant in the portal answers as that customer — their own account and published knowledge, and nothing else. This is the surface where a mistake would be most visible.
Honest refusals
Asked for something they may not see, a user is told the information exists and they do not have access, rather than told it does not exist.
Trust Centre
The rest of this area.
Security, customer data, AI permissions, audit and retention — stated plainly.
Security
Encryption, isolation, production access, credential handling.
Your customers' data
What is held about the people who contact you, and for how long.
Audit trail
Who did what, including what the automations did.
Retention
How long conversations and cases are kept.
Availability
What happens when this is the system that is down.
Take better care of every customer.
Give your team the context, knowledge and AI they need to resolve problems properly — and know who needs attention before they ask.
Keep the mailbox you already use · The AI is never metered · [email protected]