05. Custom Tools

Step-by-Step Guide: Creating and Deploying a Custom Tool for Information Extraction in Lead Indicator AI

---

Overview

This tutorial walks you through combining multiple extractions into a single custom tool. The goal is to simplify the process of collecting, storing, and leveraging user information during conversations.

---

1. Setting Up the Custom Tool

A custom tool enables your assistant to extract multiple pieces of information in one step.

Steps:

1. Navigate to Custom Tools:

   • In the dashboard, go to the Custom Tool tab and click Create a Tool.

2. Define Tool Details:

   • Unique Name: Use JSON-friendly format (e.g., extract_info).
   • Description: Clearly define the tool’s purpose. Example:
     css

     Copy code

     Use this tool to extract user information, such as name, email, address, and project details, for later reference.

3. Set Parameters: Add the required data fields with descriptions:

   • name: The user's name.
   • email: The user's email address.
   • address: The user's physical address.
   • concrete_issue: A brief description of the user's concrete-related issue.
   • concrete_details: Detailed information about the project.

4. Specify Endpoint URL:

   • This URL is a webhook to receive and process the extracted data.
   • Open a workflow tab in your CRM (e.g., Go High Level or Lead Indicator).
   • Create a new workflow and set the trigger as Inbound Webhook.
   • Copy the Webhook URL and paste it into the tool's Endpoint URL field.

5. Save and Assign Tool:

   • Save the tool and assign it to your assistant under Tools in the assistant’s settings.

---

2. Setting Up the Workflow

The workflow processes the data received from the custom tool.

Steps:

1. Create a New Workflow:

   • Name it (e.g., concrete_extraction_workflow).

2. Configure the Trigger:

   • Set the trigger as Inbound Webhook.
   • Use the webhook URL provided earlier.

3. Process the Extracted Data:

   • Add the action Find Contact:
     • Match the contact using their Phone Number from the webhook data.
   • Add the action Update Contact Fields:
     • Map the extracted parameters (name, email, address, etc.) to the corresponding contact fields.

4. Test the Workflow:

   • Use the assistant in a conversation and trigger the custom tool.
   • Check the workflow’s logs to verify that the data is processed correctly.

---

3. Testing the Custom Tool

1. Initiate a Conversation:

   • Start a chat or call with the assistant.
   • Provide sample details (e.g., name, email, address, and project information).
   • Verify that the assistant uses the custom tool to extract information.

2. Check Workflow Logs:

   • Ensure the workflow received the data and updated the contact fields.

3. Review Data Mapping:

   • Ensure the extracted data (e.g., concrete_issue, concrete_details) is stored in the correct fields.

---

4. Advanced Customizations

• Multiple Conditions: Add conditions to the workflow (e.g., specific actions based on concrete_issue).
• Dynamic Responses: Use the extracted data for tailored responses during the conversation.

---

5. Final Deployment

1. Publish All Changes:
   • Publish the assistant, tool, and workflow.
2. Automate Enrollments:
   • Add new users to workflows automatically (e.g., via CRM triggers or API calls).

---

6. Use Case Example

During a chat:

• User Input: "I need help with a 1,000-square-foot concrete foundation repair."
• Assistant Action: Extracts:
  • name: John Doe
  • email: john.doe@example.com
  • address: 123 Main St, Springfield
  • concrete_issue: Foundation repair
  • concrete_details: 1,000-square-foot residential project
• Workflow Output: Updates contact details and sends them to the sales or support team.

---

Conclusion

By consolidating extractions into one tool, you streamline data collection and processing. This approach reduces complexity, ensures consistency, and prepares your assistant for scalable deployments.