Skip to main content

​
Overview

AI disclosure appends a short sign-off to messages your AI sends on its own. Contacts see who (or what) wrote the reply without you having to add it to every agent’s instructions. The setting is off until you turn it on. Enable it in Settings → General → Features → AI disclosure.

​
What gets disclosed

  • Only messages the AI sends automatically (autopilot).
  • Only the first message of each run of AI replies. If the AI sends three messages in a row, only the first carries the sign-off.
  • A human reply resets it. The next AI message after a person replies gets the sign-off again.
  • A contact replying does not reset it. The AI is still the last one who spoke.
Example thread:

​
What does not get disclosed

  • Anything a person sends, including copilot replies. When you review and approve an AI draft, you are the sender, so no sign-off is added.
  • Attachment-only sends with no message body.

​
The sign-off text

By default the sign-off is your workspace name followed by 's AI (for example, Acme Rentals's AI).
  • Default text follows a rename. Rename the workspace and the sign-off updates with it.
  • Custom text does not. Once you type your own wording, it stays exactly as written until you change it.
  • Clear the field to go back to the default.
Custom wording has to actually disclose:
  • It must say AI or artificial intelligence.
  • It cannot deny AI authorship (This is not an AI, no AI involved, and similar are rejected).
  • It is limited to 120 characters.

​
Agent signatures

While disclosure is on, the sign-off replaces the agent’s configured signature on automatic sends, so contacts get one sign-off rather than two. Copilot drafts keep the agent signature, since a person sends those and they are not disclosed.

​
Voice

Enabling disclosure also updates your AI phone numbers. Voice agents open the call by introducing themselves as AI (“Hi, I’m Acme Rentals’s AI assistant.”), before whatever greeting you have configured. Voice always uses that spoken sentence. Your custom written sign-off is for messages only. Renaming the workspace re-pushes the sentence to every AI phone number, so agents never introduce themselves with an old company name.

​
Turning it on or off mid-conversation

The setting applies to messages sent from that point forward. Existing messages are never rewritten. Turning it on mid-thread means the next AI message is signed, even if the AI had already been replying. Earlier AI messages in that thread never carried a sign-off, so they do not count as a run that has already disclosed. Turning it off stops sign-offs immediately and restores the agent signature.