Skip to main content

Files

The Files section allows users to upload and manage documents that the chatbot can use for enhanced responses. This feature supports RAG (Retrieval-Augmented Generation), enabling the chatbot to retrieve and provide information directly from uploaded documents.

What is RAG (Retrieval-Augmented Generation)?​

RAG enables the chatbot to reference external knowledge sources such as PDFs, CSVs, and DOCX files during conversations. Instead of relying solely on pre-trained AI models, the chatbot can search through these documents to deliver more accurate and relevant responses.

Supported File Types​

  • PDF
  • CSV
  • DOCX
  • TXT

Uploading these file types allows the chatbot to access structured and unstructured information.

Use Cases​

1. Product Documentation Support​

  • Scenario: A customer asks for details about a specific feature.
  • Solution: The chatbot retrieves the relevant section from the uploaded product documentation.
  • Scenario: A chatbot for a law firm is asked about specific regulations.
  • Solution: The chatbot references uploaded legal documents to provide accurate responses.

3. Customer Knowledge Base​

  • Scenario: Users inquire about policies, services, or procedures.
  • Solution: The chatbot uses a knowledge base file (e.g., a company FAQ) to answer questions directly.

How to Use Files​

  1. Upload Files:
    Navigate to the Files section and upload your documents. Ensure the files are in supported formats.

  2. Process and Index Files:
    Once uploaded, the platform processes the files to make them searchable. This indexing step is crucial for efficient document retrieval during chatbot interactions.

  3. Enable Document Retrieval:
    Link the file data to specific chatbot tasks or enable RAG globally for the chatbot to use document references when needed.

  4. Test Document Responses:
    Use the chatbot simulator to test queries related to the uploaded files and verify that the responses are accurate and relevant.

Example Scenario​

Scenario: Customer Asks for Warranty Terms​

  1. The chatbot receives a question: "What is the warranty policy for product X?"
  2. It searches the uploaded warranty document for relevant sections.
  3. The chatbot returns a response like:
    "The warranty for product X covers defects for 12 months from the date of purchase."

Best Practices for Using Files​

  • Organize Documents: Group similar documents together for easier indexing and retrieval.
  • Use Descriptive File Names: Name files clearly (e.g., Product_Warranty_Terms.pdf) to streamline search and reference.
  • Keep Documents Up-to-Date: Regularly update files to ensure the chatbot provides current information.
  • Test Frequently: After uploading new files, test queries to confirm the chatbot retrieves the correct information.

Benefits of Using Files​

  • Enhanced Responses: The chatbot can provide detailed, document-backed answers.
  • Domain-Specific Knowledge: Upload industry-specific documents to handle niche queries effectively.
  • Improved Accuracy: Reduce errors by directly referencing authoritative documents.
  • Scalability: Manage large document collections without manually coding responses.

By leveraging the Files section, you can transform your chatbot into a highly knowledgeable assistant capable of delivering precise, data-driven responses.