1. Overview
  2. Learning the Software - Beginner
  3. 01. Knowledgebase

01. Knowledgebase

Step-by-Step Guide: Setting Up and Using the Knowledge Base in Lead Indicator Software

Purpose

The knowledge base is essential for training the Lead Indicator AI bot to handle customer inquiries accurately by providing company-specific information.

Steps for Setting Up a Knowledge Base

1. Understanding the Knowledge Base

The AI is pre-trained with industry-specific knowledge (e.g., concrete foundation repair). The knowledge base is used to add company-specific details like:

  • Services offered
  • Operating hours
  • Pricing and estimates
  • Policies (e.g., cancellation fees)

2. Accessing the Knowledge Base

  • Navigate to the left-hand menu and click on the "Knowledge" tab (represented by a stack of icons).
  • Create a new knowledge base by clicking the top-right button. Name it appropriately for organizational purposes.

3. Adding Data

There are several ways to populate the knowledge base:

A. Manual Text Input
  • Click “Add New Data Source” and select “Text.”
  • Example content:
    diff
    Apex Concrete serves Tucson, Arizona, offering: - Concrete Foundations - Driveways - Masonry Services Contact us at support@concrete.com.
  • Save the entry by naming it (e.g., “Concrete Services”).
B. FAQ Creation
  • FAQs provide precise Q&A pairs, which are ideal for improving AI accuracy.
  • To generate FAQs:
    • Visit your company’s website or use a tool like ChatGPT to create questions based on the site content.
    • Example:
      • Question: Do you offer free estimates?
      • Answer: Yes, Apex Concrete provides free estimates for all services.
  • Add FAQs to the knowledge base manually or upload in bulk.
C. File Upload
  • Upload PDFs, images, or documents containing company policies, terms, or FAQs.
  • Example: A PDF of company pricing or terms and conditions.
D. URL Integration
  • Input the URL of your company website to extract text automatically.
  • Test the extracted data for relevance.

4. Testing the Knowledge Base

  • Use the “Test Knowledge Base” feature to simulate customer queries.
  • Example queries:
    • “What services do you offer?”
    • “Do you provide free estimates?”
  • If responses are incorrect or incomplete:
    • Refine the knowledge base by editing or adding new entries.
      • Ask questions in the Test Knowledge Base section
      • If the answer is not what you want, input the correct responses
        • Example: "Query": What are your services? "Answer": What services do you offer?: We service concrete foundations, concrete driveways, concrete retainer walls, paving, and masonry work.

5. Best Practices for a Robust Knowledge Base

  • Consistency: Ensure similar data is grouped logically to avoid fragmentation.
  • FAQs First: Start with FAQs to handle specific customer questions efficiently.
  • Continuous Updates: Add new questions or modify existing entries based on customer interactions.
  • Test Regularly: Refine responses to maintain accuracy.

Troubleshooting Tips

  • If AI responses are off-target:
    • Check if the relevant data exists in the knowledge base.
    • Ensure questions and answers are specific and well-formatted.
  • For FAQs, ensure key-value pairs (question-answer) are set up correctly.
  • Use JSON formatting for complex queries if supported.
  • For a more in-depth guide on how to optimize your Knowledge Base, check out our Building and Optimizing a Knowledge Base for AI Assistants article

 

Contact Support

For additional help, reach out to support@leadindicator.ai.

Was this article helpful?