K
Kacper Lewko
HighLevel Workflows need a native, secure human review and approval step for emails generated by AI Agents.
Current problem:
An AI Agent can generate an email inside a Workflow, but when the content requires human review, there is no dedicated interface where an authorized team member can view, edit, approve, or reject the message before it is sent.
Editing multiple contact custom fields is not a practical review interface. Public forms embedded through Custom Menu Links do not provide the required access protection and contact context.
Requested functionality:
  1. Add a native “Human Review” or “Approval” action to Workflows.
  2. Allow a Workflow to pause before an AI-generated email is sent.
  3. Display the generated subject and email body in a secure editor inside HighLevel.
  4. Notify and assign the review to a selected user or team.
  5. Allow the reviewer to edit, approve, reject, or reassign the message.
  6. Keep the review page accessible only to authenticated and authorized HighLevel users.
  7. Preserve the correct contact and Workflow execution context automatically.
  8. Resume the same Workflow after approval.
  9. Provide separate paths for Approved, Rejected, Timed Out, and Reassigned states.
  10. Prevent duplicate sends and maintain an audit log showing who edited and approved the message.
  11. Support plain-text editing with optional conversion to email-safe HTML after approval.
  12. Make the approved subject and body available as values in later Workflow actions.
Example use case:
  • An AI Agent generates a personalized email.
  • The AI returns
    requires_review = true
    .
  • The Workflow pauses and creates a review request.
  • A team member receives a notification and opens the internal review screen.
  • The team member corrects and approves the email.
  • The same Workflow resumes and sends the approved message.
  • The system records who approved it and when.
Why this matters:
AI-generated communication cannot always be sent without supervision. A native approval step would allow businesses to use AI safely in sales, customer service, appointment follow-up, legal-sensitive communication, and other workflows without building external applications or unsafe public forms.
This request was discussed with HighLevel Support under ticket #6524253 and was directed to the Ideas Board by the Workflows Team.