Email autonomy modes (coming soon)
Coming soon. Choose how your agent handles email replies: in supervised mode (each reply awaits your approval) or autonomous mode (automatic sending with safety guardrails).
TWO MODES TO CHOOSE FROM
Coming soon: the email channel and its reply modes are being rolled out.
Go to 'Settings > Email > Reply mode' to choose how your agent handles email replies:
- Supervised (recommended, default): each reply is drafted by the AI then kept as a draft. Nothing is sent without your approval from the Conversations inbox.
- Autonomous: replies are sent automatically, without your intervention.
GUARDRAILS IN AUTONOMOUS MODE
Even in autonomous mode, the AI refrains from sending and holds the reply pending in three cases, for your safety:
- The reply would be empty
- The sender could not be authenticated (anti-spoofing check failure)
- The knowledge base does not contain sufficiently reliable information to reply
Frequently asked questions
WHICH MODE SHOULD I CHOOSE?
Start in supervised mode to verify the quality of replies on your first exchanges. Once you are confident, autonomous mode saves you time, while keeping the guardrails described above.
CAN THE AI SEND AN EMAIL WITHOUT MY APPROVAL?
In supervised mode (the default setting), no, never: each reply stays as a draft until you approve it from the Conversations inbox. Only autonomous mode sends automatically, and only if you have activated it yourself.
DO I NEED TO ACTIVATE A PARTICULAR MODE?
No. Supervised mode is active by default: you remain in control of every message. You only switch to autonomous mode if you decide to.