Setting Up Your Bot as a Voice Agent in Lead Indicator

Introduction

This guide shows you how to activate your AI assistant as a voice agent. You’ll learn how to connect a phone number, set up outbound calling, and carry context between SMS and calls using GoHighLevel automations and custom fields.

1. Select and Add a Voice

• Go to the top-right corner and click the Voices icon (three bot heads).
• Choose from existing voices or click Import Voice to pull one from ElevenLabs.
• Search a voice (e.g., “Zuri”), click import, and it’ll appear in your voice list.
• To clone your own voice:
  • Upload an MP3 or MP4 (under 2 minutes).
  • Go to the Cloned tab to find your custom voice.

2. Connect a Phone Number

• Click the Phone icon (top-right).
• Select an existing number or click Buy Number to purchase a new one using an area code.
• If an area code isn’t available, try another or wait until one is.
• Click the three dots on a number → choose the assistant you want to connect it to → save changes.
• The number is now live and linked to your bot for inbound calls.

3. Make an Outbound Call via Workflow

• Create a GoHighLevel workflow and use the action: Make AI Call.
• You’ll need your assistant’s ID. Go to your assistant and click the copy button next to the ID (top-left).
• Paste that ID into the workflow’s assistant ID field, save and publish.
• Add any trigger you want (e.g., form submission) to start the call.

4. Create Custom Fields for Call Context

These help maintain conversation history across SMS and voice:

• Message History – large text field
• Last Message – text field
• Last Call Transcript – large text field
• Last Call Summary – large text field
• Recording URL – text field
• Call History Summary – large text field
• Accumulated Call Time (sec & min) – numerical fields

5. Set Up a Post-Call Webhook

• Create a new workflow with an Inbound Webhook Trigger.
• Copy the webhook URL and go to your assistant’s Call Settings.
• Paste the webhook URL into the Post Call Webhook field and send a test call to generate the mapping reference.
• Back in your workflow, use the mapping reference to access call data like summaries and recordings.

6. Map Webhook Data to Contact Fields

• Find the contact using Contact ID sent from the webhook.
• Use a GPT step to create a Master Call History Summary.
• Combine the new call summary with previous summaries.
• Update contact fields:
  • Call Summary
  • Call Transcript
  • Recording URL
  • Call Time (Seconds & Minutes)

7. Carry Message History Across Channels

• Create a trigger for Customer Replied (Any Channel).
• Append the message body to Message History field.
• Update Last Message field with just the new message.
• In outbound calls, use Custom Variable 1 to pass Message History so the bot has full SMS context.

8. Add Context in Prompt Using Custom Variables

• In your outbound call workflow, pass Contact.MessageHistory into Custom Variable 1.
• In the assistant prompt, reference {{ custom1 }} to give the AI the full message context.
• Useful when switching from SMS to voice and maintaining conversation flow.

Tips

• Use short, clean names for custom fields to stay organized.
• Always test the webhook connection before finalizing the workflow.
• Use the logs inside the assistant to see how post-call data is being received.
• Clone voices under 2 minutes long for best results.

Summary

This SOP shows how to configure your bot as a voice agent by selecting a voice, assigning a number, creating outbound call workflows, and enabling post-call context syncing. You'll also learn how to build a custom field system to ensure smooth handoffs across SMS and voice channels.

• Select and add a voice (or clone your own).
• Connect a Twilio phone number to your assistant.
• Create outbound call workflows using the Make AI Call action.
• Set up a post-call webhook to capture call summaries, transcripts, and recordings.
• Use custom fields and GPT to generate context-aware history.
• Maintain message and call context across channels using Custom Variables.

Need help? Visit support.leadindicator.ai or email [email protected].