How It Works

Add AI Chat to Your Website
In Minutes, Not Months

Wisdone uses RAG (Retrieval-Augmented Generation) to create an AI chatbot that actually understands your content and answers questions accurately.

Get Started in 4 Simple Steps

1

Create Your Account

Sign up and access your dashboard. No credit card required—the free plan includes 50 messages to get you started.

2

Build Your Knowledge Base (Configure RAG)

Enter your website URL to crawl pages, or upload PDFs, Word docs, and text files. Wisdone automatically converts your content into vector embeddings and stores them in a vector store—this IS your RAG configuration. Better content = more accurate AI answers.

3

Customize Your Chatbot

Match your brand with custom colors, themes, and greetings. Configure AI behavior, response style, retrieval depth, suggested questions, and CTA buttons.

4

Embed on Your Website

Copy a single line of code and paste it on your website. Works with any platform—WordPress, Shopify, Wix, or custom HTML.

Knowledge Base

Train Your AI on Your Content

Add your website pages, documents, and FAQs. Wisdone indexes everything using vector embeddings for fast, accurate semantic search.

Website Crawling

Enter your URL and Wisdone automatically crawls, extracts content, and creates vector embeddings for your RAG pipeline.

  • Single page or recursive crawling (up to 2 levels)
  • Sitemap support for large sites
  • Automatic text extraction and chunking
  • Content hashing for efficient re-syncs

PDF Documents

Upload product manuals, whitepapers, guides, and any PDF documentation. Content is extracted and indexed in your vector store.

  • Full text extraction
  • Automatic chunking (~1000 chars)
  • Metadata preservation for citations
  • Large file support via GridFS

Word Documents

Import .docx files with your FAQs, policies, or internal documentation. Each becomes searchable in your knowledge base.

  • Rich text extraction
  • Table and list support
  • Vector embeddings generated automatically
  • Source citations link back to original

Text & CSV Files

Add plain text files, CSV data, or JSON for structured information. All content feeds into the same RAG pipeline.

  • Direct text input option
  • Spreadsheet data support
  • Immediate vector indexing
  • Flexible formatting
Features

Powerful AI Chatbot Features

Everything you need to deliver exceptional customer support and capture leads.

RAG-Powered AI Responses

Retrieval-Augmented Generation ensures every answer is grounded in your actual content, not hallucinations. Your knowledge base IS your RAG configuration.

  • Automatic vector embedding generation
  • Fast semantic search using vector store
  • Configurable retrieval depth (1-20 documents)
  • Query rewriting for better context understanding

Source Citations

Build trust with visitors by showing exactly where information comes from.

  • Clickable links to original sources
  • Page titles and file names included
  • Toggle citations on/off per preference
  • Supports both web pages and uploaded files

Intelligent Lead Capture

Collect contact information naturally during conversations without annoying pop-ups.

  • Capture email, phone, name, or custom fields
  • AI decides optimal timing for capture prompts
  • Configurable trigger conditions (keywords, message count)
  • Passive, proactive, or aggressive capture modes

Real-Time Notifications

Never miss a lead with instant email alerts when visitors submit contact information.

  • Instant email notifications
  • Multiple recipient support
  • Includes conversation context and summary
  • Configurable notification preferences

Smart Call-to-Action Buttons

Convert conversations into actions with contextual CTA buttons.

  • Dynamic CTAs triggered by conversation context
  • Support for URLs, emails, phone calls, webhooks
  • Custom button styles and icons
  • Multiple display modes (after greeting, every response, end of chat)

Full Customization

Make the chatbot truly yours with extensive appearance and behavior options.

  • Custom colors, themes, and glassmorphism effects
  • Configurable AI persona and response style
  • Custom greeting messages and suggested questions
  • Widget button size, animation, and display delay

Automatic Content Sync

Keep your knowledge base and vector store up-to-date with scheduled synchronization.

  • Daily or weekly sync schedules
  • Smart change detection (only re-index modified content)
  • Automatic vector embedding updates
  • Sync status tracking and error reporting
  • Manual sync option available anytime

Analytics & Insights

Understand what visitors are asking and optimize your content.

  • Message and conversation tracking
  • Lead capture analytics
  • Usage monitoring and limits
  • Export data for further analysis

Human Handoff

Escalate to your team when a visitor needs a real person or the AI can't resolve a request.

  • Triggers on user request, low confidence, or repeated failures
  • Route to email, Slack, or a webhook
  • Reply from Slack—messages relay back to the visitor live
  • Business-hours awareness for after-hours handling

Integrations

Connect Wisdone to the tools you already use.

  • Slack for live team handoff
  • Webhooks from CTA buttons or escalations (Zapier, Make, CRM)
  • Export leads for your CRM or email tools
Under the Hood

How Your Content Becomes Searchable

When you add content to your knowledge base, it goes through our RAG pipeline. This is the same as "configuring RAG"—there's no separate setup needed.

1

Content Ingestion

Your content (web pages, PDFs, docs) is extracted and cleaned

2

Chunking

Text is split into ~1000-character segments with overlap for context

3

Embedding

Each chunk is converted to a vector that captures its semantic meaning

4

Storage

Vectors are stored in the vector store with metadata (source URL, title)

5

Query Processing

Visitor questions are embedded and matched against your vectors

6

Response Generation

Top matching chunks are sent to the LLM to generate grounded answers

Key insight: Configuring RAG = configuring your knowledge base. Better, more comprehensive content leads to more accurate AI responses. The vector store is managed automatically—you just focus on your content.

Technology

How RAG Answers Questions

When a visitor asks a question, Retrieval-Augmented Generation combines semantic search with language models to deliver accurate, grounded responses.

1. Visitor Asks a Question

A visitor types a question in your chatbot widget.

2. Semantic Search

The question is converted to a vector embedding and matched against your knowledge base using semantic search.

3. Retrieve Relevant Content

The most relevant chunks from your content are retrieved (configurable 1-20 documents).

4. Generate Response

The AI model (OpenAI GPT or Anthropic Claude) generates an accurate answer using the retrieved context.

5. Response with Citations

The visitor receives a helpful answer with clickable source citations.

Compatibility

Works with Any Website

Just one line of code. Compatible with all major platforms and custom websites.

WordPress
Shopify
Wix
Squarespace
Custom HTML
Any CMS
Embed Code
<script id="wisdone" src="https://wisdone.ai/widget.js?c=YOUR_ID"></script>

Ready to Add AI to Your Website?

Start for free today. No credit card required.

Get Started Free Read FAQ