VoiceVision - RAG-Enabled AI Assistant

VoiceVision - RAG-Enabled AI Assistant

Welcome to VoiceVision! In this tutorial, we’ll guide you through the step-by-step process of creating a RAG-enabled AI assistant. By the end of this tutorial, you'll have a fully functional AI assistant that can retrieve real-time knowledge and provide intelligent responses. Let’s get started!

Step 1: Setting Up Your Knowledge Storage

Before creating your assistant, you need to add your knowledge files to the Knowledge Storage. VoiceVision provides two options for adding knowledge:

Option 1: Importing Website Data

This method allows you to crawl content from a website and use it as a knowledge base.

Navigate to the Knowledge Storage section.

Click the “Import Website” button.

Enter a Display Name for easy reference.

Enter the Crawling URL – this is the starting point for gathering data (must begin with https://).

Enter the URL Match – this ensures the crawler fetches only relevant content. If you want to crawl everything, set this value the same as the Crawling URL.

Enter the Selector Value – this specifies which part of the webpage content should be extracted.

Set the Number of Pages to Crawl – choose a value between 1 and 200.

Once all the settings are configured, click “Create” to start crawling.

✅ Congratulations! You have successfully imported website data into your knowledge storage.

Option 2: Uploading Files

This method allows you to upload files such as PDFs, DOCs, TXTs, and other supported formats.

Click the “Upload File” button.

Select the knowledge file you want to upload.

Click “Upload” to start importing the file.

✅ Now, your knowledge storage is set up with both website data and uploaded files!

Step 2: Creating a RAG-Enabled AI Assistant

Now that your knowledge storage is ready, it's time to create your RAG-enabled AI assistant.

Navigate to the Assistants page.

Click the “New Assistant” button.

You’ll see two assistant creation options:

Simple Assistant – Works with an instruction prompt without external knowledge.

RAG-Enabled Assistant – Uses external knowledge sources such as files and website data for smarter AI responses.

Select RAG-Enabled Assistant to proceed.

Step 3: Configuring Your AI Assistant

Enter your assistant’s Display Name and Welcome Message to greet users.

Enter an Instruction Prompt that defines your assistant’s purpose and tasks.

📌 Before entering your instruction prompt, we highly recommend referring to our guideline document to craft an efficient prompt.

Select the Knowledge Documents you want your assistant to reference.

You can edit or update this list anytime after creation.

Choose an AI Model.

We recommend using GPT-4o-mini for optimal performance.

If you want voice interaction, enable the Voice Option and select a Voice Name from the list.

Step 4: Finalizing and Testing

Click the Save button to create your assistant.

🎉 Congratulations! You have successfully created a RAG-enabled AI assistant.

Step 5: Testing Your AI Assistant

Now that your assistant is set up, it's time to test it.

Click the Chat Icon on the assistant page.

Start interacting with your AI assistant to see how it responds using the knowledge data.

That’s it! You’ve now built a powerful AI assistant with real-time knowledge retrieval, enabling seamless interactions and intelligent responses. 🎯

Sample instruction prompt

You are an AI Assistant for [your-website-name-here] visitors. Your role is to assist users with the platform’s features, functionality, and usage based on the provided document information.

Response Limit: Keep all responses concise, with a maximum of 60 words.

No Additional Information Requests: Use only the given attached document information; do not ask for extra details.

Stay on Topic: If a user asks about anything unrelated to [your-website-name-here], do not answer. Instead, refocus on platform assistance.