---
doc_version: "1.0.0"
last_updated: "2026-07-14T15:00:57.718Z"
title: "What Are Knowledge Gaps?"
description: "Understand how ChatbotIQ detects knowledge gaps (topics where users ask questions but your knowledge base can't provide good answers) and what to do about them."
keywords: [knowledge gaps, gap detection, confidence scores, content gaps, analytics, ChatbotIQ]
---

A knowledge gap is a topic where users are asking questions but your knowledge base doesn't have good answers. ChatbotIQ detects these automatically and tells you exactly what content to add.

---

## How gap detection works

Every time your bot answers a question, ChatbotIQ tracks how well it was able to answer, from strong matches to no relevant content found at all.

Periodically, ChatbotIQ analyzes all the poorly-answered queries and groups related questions into **topic clusters**. Each cluster represents a knowledge gap, an area where your content is missing or insufficient.

---

## Types of knowledge gaps

ChatbotIQ classifies each gap into one of three types:

### Content gap
Your product or service covers this topic, but your knowledge base is missing documentation about it. **Action: add the missing content.**

For example, if users keep asking "How do I export data?" but you haven't documented the export feature, that's a content gap.

### Product gap
Users are asking about a feature or capability that your product doesn't have. **Action: note as product feedback.**

For example, if users ask "Can I integrate with Slack?" and you don't offer Slack integration, that's a product gap. You might add a Q&A entry explaining this, or share the feedback with your product team.

### Out of scope
The question isn't related to your product or bot's purpose. **Action: consider refining your bot's system prompt to explain its scope.**

For example, if users ask "What's the weather today?" to your documentation bot, that's out of scope.

---

## What makes gap detection valuable

### No guesswork
Instead of wondering what content to add next, you have a data-driven list of topics sorted by how many users asked about them.

### Real user language
The sample queries show you the exact words users typed. This helps you write content that matches how people actually search, not just how your team would phrase things.

### Prioritization
Gaps are ranked by query count and decline rate. A gap that 50 users hit is more urgent than one that 3 users hit.

### Progress tracking
You can mark gaps as resolved to track your improvement over time.

---

## The improvement cycle

1. **Launch your bot** and let it handle real conversations.
2. **Check Analytics** after a week or two. Look at the knowledge gaps section.
3. **Address content gaps** by adding or updating your knowledge sources.
4. **Refresh your sources** so the bot picks up the new content.
5. **Check again** in another week. The gaps should shrink.

Over time, your bot gets smarter, not because the AI model improves, but because your knowledge base covers more of what users actually ask about.

---

## Related

- [Find and Fix Knowledge Gaps](/how-to/find-knowledge-gaps/) - practical guide to using gap detection
- [How RAG Works](/explanation/how-rag-works/) - understand why gaps happen
- [Add PDFs and Q&A Sources](/tutorials/add-pdf-and-qa-sources/) - quick way to fill gaps

## Sitemap

Full documentation index: [sitemap.md](/sitemap.md) and [llms.txt](/llms.txt).