Tools vs Automatic Extractions

Tools & Abilities vs. Automatic Extractions: Which Should You Use?

Choosing between Tools & Abilities (regular tools) and Automatic Extractions is key to optimizing your AI assistant’s speed, consistency, and data handling. Understanding the difference helps you build smarter flows—reducing latency and maximizing value from each conversation.


Tools & Abilities: Real-Time Actions

  • What Are They?
    These are the tools you add in your AI editor’s toolkit (under “Tools & Abilities” or “Manage Tools”).
    Examples: book appointments, update user details, transfer call, add/remove tag, fetch from APIs.

  • When to Use:
    Use regular tools when you need something to happen during the live conversation (e.g., schedule appointments, transfer a call, update user details in real-time).

  • Tradeoff:
    The more tools you add, the more latency you introduce. If your assistant has 5+ real-time tools, expect slightly slower voice response times.

🛠️ Tip: Use regular tools for any action that must happen while the user is still on the call or in the chat.


Automatic Extractions: Post-Conversation Data Collection

  • What Are They?
    Found at the bottom of your toolkit (“Automatic Data Collection” or “Automatic Extractions”).
    These extract data after a call or chat has ended.

  • When to Use:
    For details that don’t need to affect the live conversation—such as analytics, call summaries, identifying client intent, collecting details for notifications/workflows.

  • Advantages:

    • No latency added during the call/chat.

    • AI can respond more quickly, as there are fewer in-the-moment tools to process.

  • Examples:
    Use to summarize call outcomes, collect interest type (like car model/color), or extract names/emails for post-call follow-up automations.

Note: If you only need information after the conversation (for reporting or workflow triggers), automatic extractions are often the better choice!


Choosing the Right Approach

Ask yourself:

  • Does this action need to happen while I’m talking to the client?
    → Use Tools & Abilities

  • Can I wait until after the conversation to extract this info?
    → Use Automatic Extractions

Typical use cases:

  • Live actions: Booking, transfers, updating details, custom API calls.

  • Post-conversation: Analytics, call summaries, extracting preferences/interests, triggering follow-up workflows.

💡 Pro Tip: You can combine both approaches—use tools for what needs to happen live, and extractions for everything else. This gives you fast, responsive conversations and rich analytics with minimal delay.

 

 


Was this article helpful?
© 2025 Nexis AI