Best Practices for Knowledge Bases
Organize and maintain your document libraries for better AI accuracy
How you structure your knowledge base has a direct impact on how accurately the AI answers questions. These practices consistently produce better results.
Organize by topic, not by file type
Create separate knowledge bases for distinct subject areas — "Product FAQs," "Pricing & Plans," and "Return Policy" rather than one large "Everything" library. Focused libraries give the AI a smaller, more relevant pool to search, which improves precision.
Keep individual documents focused
A 10-page document covering one topic retrieves more accurately than a 100-page document covering ten. If you have a large policy manual, consider splitting it into individual topic files before uploading.
Match the similarity threshold to your content
The default threshold of 0.7 works well for clearly written, topic-specific documents. Lower it (e.g. 0.6) if the AI says it doesn't have information when it should. Raise it (e.g. 0.8) if responses feel off-topic or vague.
Keep documents current
The AI answers based on what is in the knowledge base — not what is on your live website. When prices, policies, or product details change, re-upload the relevant document so responses stay accurate.
Test after every change
After adding, updating, or removing a document, send a few test questions through your support chat to verify the AI is using the updated content before sharing it with customers.
What's next
- Document and File Size Limits — check your plan's upload limits
- How to Link a Knowledge Base to a Workflow — make your library active for a support chat
Was this article helpful?
Build AI chatbots like a pro
Take control of your customer support with Voxe. Create unlimited AI-powered chatbots, manage knowledge bases, and automate workflows—all with complete data ownership.
Start your 14-day free trial • No credit card required