Trust & security
AI governance
AI acts only where the workflow says it can. People make the decisions that matter.
Our principles
- Nothing is sent or posted without your OK when you set it that way. In Social, every post waits for approval.
- Knowledge answers come from the documents you approve, and show where the answer came from.
- Access, data handling and approved knowledge sources are agreed at the start, not added later.
- Results are measured against a baseline.
- Your team is trained to use, check and improve what is deployed.
In Base
- Bots can only send on their own within limits their maker sets.
- If someone else changes a bot, its sending on its own pauses until its maker turns it back on.
- Every connector call is recorded.
Have a security or data question?
Email [email protected] and we will get back to you.