Knowledge Base Management
Knowledge Base Management

Master the art of AI training. Learn how to build an expert Knowledge Base for your Passerbot agents using document indexing and automated website crawling.

Last updated on Mar 7, 2026

Knowledge Base: Building the Brain of Your AI

The Expertise and Authoritativeness of an AI agent are directly tied to the quality of the data it consumes. In the Passerbot ecosystem, the Knowledge Base is the high-fidelity repository of your business intelligence.

By centralizing your product manuals, FAQs, and technical documentation, you empower your AI Agents to provide responses that aren't just intelligent, but are ground-truth accurate for your specific business.


Strategies for Expert-Level Training

A well-architected Knowledge Base is the difference between a generic chatbot and a specialized business assistant. We recommend a multi-source approach:

1. High-Fidelity Document Indexing

For deeply technical or structured information, document uploads provide the highest level of detail. Passerbot’s parsing engine is optimized for:

  • Project Manuals & Guides (PDF): Perfect for comprehensive product documentation where formatting matters.
  • Internal Knowledge Docs (DOCX): Ideal for training agents on company policies and "tone of voice" guidelines.
  • Structured Data (XLSX): If your AI needs to reference price lists, product SKU details, or inventory specs, Excel indexing provides the structure it needs for precise lookups.

2. Automated Website Sync (Crawling)

Your website is often the most up-to-date source of truth. Instead of manually re-uploading files every time you update a page, use our Sitemap Crawler:

  • Point and Click: Provide your sitemap URL (e.g., yourdomain.com/sitemap.xml).
  • Deep Indexing: Our system explores your public pages and extracts relevant text content, creating an authoritative digital twin of your website's expertise.
  • Synchronization: Keep your AI "in the loop" with your latest blog posts, product launches, and company news.

Architecture: Segmenting Your Knowledge

To maintain high trustworthiness, it is often necessary to limit what an AI knows based on its role. Passerbot supports Knowledge Segmentation:

Scenario: The Sales vs. Support Split

You might create two distinct Knowledge Bases:

  • "Public Marketing" Base: Contains pricing, feature lists, and case studies. Linked to your Sales Agent.
  • "Technical Support" Base: Contains troubleshooting steps, API references, and internal guides. Linked to your Support Agent.

By segmenting knowledge, you prevent "hallucination" and ensure that each agent remains a focused expert in its specific domain.


Maximizing Trust Through Accuracy

A Knowledge Base isn't just a container; it's a commitment to your customers that the information they receive is sanctioned by your business.

  • Accurate Citations: Because our agents reference your files directly, they are less prone to the creative errors found in general-purpose AI.
  • Easy Updates: Need to change a policy? Simply replace the file in your Knowledge Base, and all linked AI Agents are instantly updated.

Start building your foundation today by following our Quick Start Guide to initialize your first knowledge segment.

?Frequently Asked Questions

Q:What file formats are supported for the Knowledge Base?

Passerbot supports professional document formats including PDF (ideal for manuals), Word (DOCX), Excel (XLSX for structured data), and Plain Text (TXT).

Q:How does website crawling work?

You provide your website's sitemap URL, and Passerbot's automated crawlers index the text content of your pages. This ensures your AI stays updated with your latest website content without manual file uploads.

Q:Can I have separate knowledge bases for different agents?

Yes. You can architect multiple knowledge bases and selectively link them to specific agents. This ensures, for example, that your Sales Agent only accesses marketing data while your Support Agent accesses technical manuals.