Saltar al contenido principal

Knowledge Base Overview

info

AI SmartTalk's Knowledge Base enables seamless chatbot enrichment through diverse file imports. This overview outlines the types of supported files and their practical applications.

Upload form

The Knowledge Base Interface is the primary management console for your Retrieval-Augmented Generation (RAG) system. It lets you browse, filter, classify and maintain the FAQ entries and other content that your chatbot retrieves at runtime.

Knowledge base page

🔍 Filters​

Filters

Click Filters to expand or collapse the filter panel. Use these controls to narrow the list of entries:

  • Status
    Select lifecycle stages (e.g. Active, Draft, Archived)
  • Category
    Choose from your taxonomy (e.g. Billing, Technical, Other – To Classify)
  • Source
    Filter by origin (e.g. FAQ, Help Center, Docs)
  • Clear All
    Resets every filter back to its default state.

🛠️ Batch Actions​

Batch Actions

Use the Batch actions dropdown to perform bulk operations on selected entries:

  • Enable / Disable
  • Change Status (e.g. Archive, Publish)
  • Reassign Category
  • Delete

📋 Custom Columns​

You can toggle additional columns on or off to suit your workflow, such as:

  • Last Updated
  • Author
  • Last Used

📑 Entry Table​

ColumnDescription
NameThe title or first line of the FAQ/knowledge item (click to edit).
UsageHow many times this entry has been retrieved by the chatbot in live chats.
StatusWhether the entry is Active, Draft, or Archived.
CategoryA high-level tag (e.g. “Billing,” “Technical,” “Other – To Classify”) used for routing and analytics.
SourceOrigin of the content (e.g. “FAQ”, “Help Center”, “Docs”).
ActionsContext menu (…) to edit, preview, delete or view history for one item.

🔄 Pagination​

  • Navigate through pages of entries using the Prev/Next buttons or jump directly to a page number.

Pagination

🚀 How It Powers the Chatbot​

  1. Retrieval
    When a user asks a question, the chatbot queries this Knowledge Base for the most relevant entries.
  2. Augmentation
    Retrieved content is passed into the language model’s prompt to ground its responses in your official documentation.
  3. Feedback Loop
    Usage metrics (the Usage column) help you identify high-value entries and gaps in coverage, so you can continually refine your RAG corpus.

Supported Import Types​

FAQ Files (Excel/CSV)​

Upload structured question-and-answer data in Excel or CSV formats to enhance chatbot response accuracy.

  • Best for: Frequently Asked Questions (FAQs) and pre-defined customer queries.
  • Example Format:
    • Column 1: Questions
    • Column 2: Answers
    • Column 3 (Optional): Links or additional context.

Details: Import FAQs with Excel or CSV

Document Files (DOC, DOCX, PDF)​

Import detailed documents to generate searchable knowledge chunks automatically.

  • Best for: Company manuals, in-depth guides, or training materials.
  • Processing: Smart chunking for efficient chatbot integration.

Details: Import Document Files

Google Merchant Store Files​

Integrate e-commerce product feeds directly into the chatbot knowledge base, ensuring up-to-date information.

  • Best for: Product catalogs and inventory management.
  • Requirement: A unique "product-id" column for tracking and updates.

Details: Google Merchant Store Integration

tip

For better chatbot responses, ensure imported data is clean, categorized, and tagged appropriately.