Website | A.I. Voice Recordings
Processes Voice OS call webhooks: matches the caller, logs the AI call, updates fields, then branches on answered/not-answered with DND handling.
When an AI voice call is made or received through the partner channel, this automation fires via webhook. It finds the contact, logs a call summary note, extracts and updates key information from the conversation into the contact record, then branches based on whether the call was answered or not. Answered calls are tagged accordingly and checked for DND (Do Not Disturb) status โ enabling or disabling DND based on the contact's preference. Unanswered calls are handled in a separate branch and exited cleanly.
Every AI voice interaction is automatically logged, summarised, and acted on โ no manual data entry, no missed follow-up flags, no compliance gaps. The DND logic is particularly important: it ensures the system respects contact preferences automatically, protecting both the client's reputation and their legal obligations around outbound communication. The call summary note means the team always has context on every voice interaction without needing to listen back to recordings.
This is the operational backbone behind the AI Voice Agent in the partner channel. It's what makes voice AI actually useful in a business context โ the call doesn't just happen and disappear, it feeds structured data back into the CRM and triggers the right next action. In sales conversations, this is the answer to 'what happens after the AI makes a call?' โ everything is captured, logged, and acted on automatically.
This workflow processes inbound Voice OS website deployment webhook events, identifies the correct contact using their phone number, logs the AI call details, updates call-related custom fields, and branches based on whether the AI call was answered. If answered, it further checks whether the contact should be placed on Do Not Disturb (DND) for calls, manages tags, updates DND settings, sends Slack notifications, and removes the contact from other calling workflows where appropriate.
The build โ trigger โ steps โ branches
Voice OS | Website Deployment Webhook
Inbound Webhook
Fires when a webhook is received from Voice OS / Website Deployment. Expected data: lead phone number, call transcript, call duration, call status, recording URL, AI-extracted callback time, AI-extracted notes (used for DND detection).
Find Contact
ActionSearch by Phone ยท {{lead.phone_number}}
Locate the existing contact so all following actions are performed on the correct record.
Add Contact Note
ActionLog call summary
Creates a formatted note containing call transcript, call duration, call status and recording URL; maintains a complete call history on the contact timeline.
Update Contact
ActionVoice OS | Call Status = Call Status ยท Voice OS | Call Back Time = AI Extracted Callback Time
Stores AI-generated call info (e.g. Complete / Failed / Busy / No Answer) for future automation and reporting.
Was the call answered? (Voice OS | Call Status contains 'complete')
Call Status contains 'complete'.
Add Tag
Actionai call | answered
Identify contacts who successfully answered an AI voice call.
Check DND Status
If / ElseAI Extracted Note compared against 'DND Extracted'
Branches on whether the contact requested Do Not Disturb.
Enable Call DND (if DND)
ActionChannel: Calls Only
Prevent future AI phone calls while allowing other communication channels.
Slack Notification (if DND)
NotifyTeam alert: contact placed on Call DND
Remove From Calling Workflows (if DND)
ActionRemove from active AI calling sequences
Prevent future call attempts.
Slack Notification (if no DND)
NotifyTeam alert: call answered, no DND detected
Call Status does not contain 'complete'.
Add Tag
Actionai call | not answered
Identify contacts who missed or did not answer the AI voice call for future follow-up campaigns.


