Knowledge Base

The knowledge base is what your AI agent searches to answer customer questions. The more relevant knowledge you add, the better your agent's responses will be.

How it works

1. You add content (URLs, PDFs, text)
2. SuperWaba splits the content into chunks and generates vector embeddings
3. When a customer asks a question, the agent searches for the most relevant chunks
4. The agent uses those chunks to generate an accurate answer

Adding knowledge

From a URL

1. In the Knowledge Base tab, click Add from URL
2. Enter the URL (e.g., your FAQ page, product page, or blog post)
3. SuperWaba crawls the page, extracts the text content, and indexes it

You can add multiple URLs. Each page becomes a separate knowledge document.

From a PDF

1. Click Upload PDF
2. Select the file (max 10 MB)
3. The PDF is parsed, text is extracted, and indexed

From text

1. Click Add Text
2. Enter a title and paste the content
3. Save — the text is chunked and indexed immediately

Best practices

• Be specific — Add content that directly answers the questions your customers ask
• Keep it current — Update knowledge docs when your policies or products change
• Quality over quantity — 10 well-written FAQ entries are better than 100 pages of irrelevant content
• Use clear titles — Titles help the agent understand what each document is about

Per-agent knowledge

Each agent can have its own set of knowledge documents. This lets you:

• Give your sales agent product knowledge
• Give your support agent troubleshooting guides
• Keep sensitive information restricted to specific agents

Chunk size

Content is automatically split into chunks of ~500 tokens. This is optimized for retrieval accuracy — you don't need to configure this.