AI and agents
Can the agent publish without a human?
No. The BetterCMS agent never publishes on its own. Every change it proposes waits for a person to approve it, and each field shows its before and after so the approval is informed rather than one blanket yes. Undo is available afterwards as well, per document.
The rule
The agent proposes. A person approves. There is no configuration that changes this, and no plan on which it behaves differently.
Why a rule and not a setting
Because a setting gets turned on. It gets turned on by somebody in a hurry on a Thursday, and the failure it enables is not a typo, it is a confident, plausible, wrong change applied to forty documents while nobody is watching.
The review is not a safety feature bolted onto the agent. It is the product. An agent you have to check is worth having; an agent nobody checks is a liability with a subscription.
What it costs you
Time, honestly. Somebody has to look at the diff, and on a large run that is a few minutes rather than none. That is the trade, stated plainly, and it is the right one for a system whose output is public.
What it saves is the other few minutes: the ones spent finding out which of forty documents went wrong, three weeks later, from a customer.
Also in ai and agents
Still have a question left
Thirty minutes with the product team, and a straight answer.