Scoped access, a named approver, everything logged.
What Site Brain can touch, who has to say yes, what is recorded, and what happens to your data if you stop. Each piece is below, in full.
Each piece in full
Access
Read-only by default. Write access only through the connector you chose, with the narrowest scope it allows. The agents don't act outside that scope.
Approvers
A named person on your side approves. You can set standing approvals for low-risk changes and switch them off again, and the agents don't approve their own work.
The log
A log of each proposal, approval, change and reversal, with who did it and when.
Rollback
Each change records what it replaced and can be put back, which matters because any change can break something. When a check fails, it stops and asks.
Credentials
Credentials sit in a vault and are used per task, not stored where an agent can read them.
Your data
Read access to your analytics, search and ad accounts. Disconnect and the data is deleted. Changes made through your CMS or your code stay on your site. Edge changes last as long as the connector runs unless they've been written back to the source. The history is yours to export.
Questions for your security team?
Email me and I'll answer them. I reply within a business day, myself.
Email me