Automatic Data Collection Tools (Automatic Extractions)

Automatic data collection tools—also known as automatic extractions—are a game-changer for extracting valuable information from your AI assistant’s call transcripts or chat conversations after the interaction is complete. These tools allow you to capture structured data such as customer intent, specific details mentioned in the conversation, or any other custom value you want to track—without interrupting the flow of the live interaction.

---

What Are Automatic Data Collection Tools?

Automatic extractions work by scanning the call or chat transcript and pulling out the information you specify. Unlike regular tools (which work during the conversation), automatic extractions run after the conversation ends. This makes them ideal for capturing important details for CRM updates, analytics, or follow-up automations—without slowing down the live experience.

When Should You Use Them?

• During the call: Use regular tools from the toolkit (Tools & Abilities).
  
  

• After the call/chat: Use automatic data collection tools for post-call or post-chat data extraction.

💡 Tip: If you want to extract information from a call transcript (like caller intent or specific details), set up automatic extractions. For chat, you can also use these, but if you need real-time actions, use regular tools.

---

Use Cases

• Industry-specific summaries: e.g., extracting call purpose for reporting or analytics.
  
  

• Filling CRM fields: e.g., auto-capturing “service needed,” “budget,” or “product interest.”
  
  

• Custom follow-ups: Use extracted values to trigger workflows or send targeted follow-up emails/texts.

---

How to Set Up Automatic Data Collection Tools

1. Create Custom Fields in Your CRM

• Go to your CRM settings.
  
  

• Add any new custom fields you want to capture (e.g., “Caller Intent,” “Shirt Color”).

2. Resync Fields (If Needed)

• If you create a new custom field after starting, use Fields & Values > Resync in your AI editor. This ensures new fields are available for mapping.

3. Configure Automatic Data Extractions in Your AI Editor

• Go to your AI assistant editor.
  
  

• Open Toolkit > Automatic Data Collection.
  
  

• Click New Data Extraction.
  
  

  • Name it (e.g., “Caller Intent”).
    
    

  • Choose the data type (String, Boolean, Number).
    
    

  • Write a description that tells the AI what to extract (e.g., “Summarize the caller’s purpose for the call in one phrase”).
    
    

  • Map the extraction to the correct custom field in your CRM.
    
    

4. (Optional step for more reliability) Connect with Post Call Webhook (for Voice/Calls)

• For calls (and voice widgets), use a post call webhook to send all call data—including extractions—to your automation system or CRM.
  
  

• In your CRM automation (e.g., via inbound webhook trigger), map the extracted data from the webhook payload to the appropriate custom fields.

5. Troubleshooting

• Sometimes data may not appear immediately in the custom fields. Double-check the mapping and ensure you’re using the latest payload in your automation.
  
  

• For chats, regular tools may sometimes be a better option, since latency isn’t a major issue in chat as it is in voice.
  
  

⚡ Secret Sauce: Automatic extractions are typically a better choice for pulling info after calls, since they don’t add any latency to live conversations and don’t overload your AI with too many real-time tools.