How to Set Up Team Notifications for AI Chat Responses

If your AI is handling chat and text messages, it’s crucial to keep your team in the loop—without overwhelming them with constant notifications. This guide shows you how to set up automated notifications summarizing AI activity, so your team always knows what’s happening in every conversation, while avoiding notification overload.

 

---

 

Why Use Summary Notifications?

Instead of pinging your team every time a new message comes in (which gets overwhelming fast), set up a system to notify them only after conversations are completed. This lets your team track AI responses and conversation outcomes without missing key details or feeling swamped.

🔔 Tip: Use summary notifications to streamline your team’s workflow and maintain awareness of all client conversations across SMS, Facebook Messenger, Instagram, and more.

 

---

 

How to Set Up Summarized Chat Notifications

Step 1: Two-Workflow System

You’ll need two workflows:

1. Message Workflow: Detects new messages across all channels.
   
   

2. Summary & Notification Workflow: Collects conversation logs, summarizes them, and sends notifications.
   
   

If you already have a workflow that triggers on every new message, you can simply add the actions described below.

Step 2: Building the Message Workflow

• Trigger: Customer reply (filter if you want to omit any channel)
  
  

• Actions:
  
  

  • Remove from Summary Workflow: Temporarily removes contact from the summary workflow to reset the timer.
    
    

  • Add to Summary Workflow: Immediately adds them back, restarting the wait timer.
    This ensures notifications are only sent after your chosen time interval (not after every message).
    
    

⏱️ Note: This timer reset is the secret sauce for controlling notification frequency!

Step 3: Building the Summary & Notification Workflow

• No trigger needed—contacts are added from the first workflow.
  
  

• Wait: Set the wait time (e.g., 10, 20 or more minutes).
  
  

• Custom Webhook:
  
  

  • Use a “conversation logs” endpoint to pull recent chat history.
    
    

  • Make sure your API key and endpoint details are accurate—small mistakes (like extra spaces) will break the connection.
    
    

• GPT Summarization:
  
  

  • Use a GPT action to summarize the pulled conversation logs.
    
    

  • Choose a model (e.g., GPT-4o mini) for best cost-effectiveness.
    
    

  • Add a custom prompt to ensure the summary covers key points.
    
    

• Internal Notification:
  
  

  • Set up an internal notification (email, SMS, etc.) with the summary included.
    
    

  • Adjust notification content to your team’s needs.
    
    

Test with a contact who has conversation history to ensure it works. Adjust settings, prompts, and timing as needed.

💡 Pro Tip: The custom webhook step uses API keys from your agency account (found in Integrations > API Keys, version 1). Protect your key—never share it!

[ POST http://createassistants.com/api/1.1/wf/get-ai-conversation

params:

location_id - string

contact_id - string

Authorization: Bearer <assistable_v1_api_key>

start_date - date (optional)

RETURNS:

{"messages":[],"returned_an_error":false,"error_message":"null"} ]

 

---

 

Troubleshooting

• If notifications aren’t coming through, double-check:
  
  

  • The API key and header formatting (no extra spaces!)
    
    

  • Workflow timing and triggers
    
    

  • That the test contact has active conversations