# Bot’s Knowledge

***

## Overview

The **Knowledge Base** feature, located in the **Bots** module, is where you teach your bot to understand and accurately answer customer questions. It helps you build a structured and smart AI assistant by combining:

* Pre-defined question–answer entries
* Business-specific terminology via Dictionary
* Auto-detection of unanswered queries
* Suggestions from real conversations with agents

> This module improves continuously as your bot interacts with users and learns from experience.

***

## 📘 1. Knowledge Entries

### Overview

This is the main list of Q\&A pairs that the bot uses to provide automatic replies. Each entry contains a question, an answer, and a status label.

### Status Types

* **Imported** – Already learned and usable
* **In Review** – Pending admin approval
* **Overlapped** – Similar to an existing entry

### How to Use

1. Go to **Bots → Knowledge Base**
2. View the list of Q\&A entries
3. Click ✏️ to edit or 🗑️ to delete any entry

> You can search entries using the top-right bar.

![Figure 1. Bot’s knowledge section](/files/JwDgIlZTtS3gSeflDlRh)

***

## 2. Dictionary

### Overview

The **Dictionary** is used to help the bot understand **slang, abbreviations, or business-specific terms**. It allows you to define what each keyword means so the bot can interpret customer input more accurately.

> ❗ The dictionary only supports plain text (no emojis or symbols).

### How to Use

1. Click **Dictionary** from the top-right corner of Knowledge Base
2. Click **+ New Keyword**
3. Enter:
   * **Keyword** (e.g., “promo”)
   * **Meaning** (e.g., “discount”)
4. Click **Save**

<figure><img src="/files/E3MIdovkKfHJP5Hmaiwm" alt=""><figcaption><p>Figure 2. Dictionary Panel</p></figcaption></figure>

***

## 3. Unanswered Questions

### Overview

When the bot is asked something it doesn't understand (i.e., not in the trained knowledge base), the question is logged as **Unanswered**. This lets admins review and decide whether to create a new answer.

> ⚙️ You can configure how the bot handles these events in Bots → Settings.

### How to Use

1. Go to **Bots → Knowledge Base**
2. Click **Waitlist → Unanswered Questions**
3. Use:
   * ✏️ to add a new answer and train the bot
   * 🗑️ to delete irrelevant/spam entries

<figure><img src="/files/3YTZTriHA5M0ZuK7Fi7p" alt=""><figcaption><p>Figure 3. Unanswer Question Panel</p></figcaption></figure>

***

## 4. Learn from Conversation (via AI Button)

### Overview

This feature allows the bot to **learn Q\&A pairs from real conversations**, but only when agents manually trigger learning via the **AI button in the bottom-left of the chat window**.

> ⚙️ Must be enabled in Bots → Settings under “Bot learns from conversation”.

### How it Works

* When the AI button is clicked during a live chat, the system will extract the most recent question–answer exchange.
* These entries are stored in **Waitlist → Learn from Conversation** for admin review.

### How to Use

1. Enable the feature in **Bots → Settings**
2. During chat, agent clicks the **AI button**
3. Go to **Bots → Knowledge Base → Learn from Conversation**
4. Use:
   * 📥 Import to accept
   * ✏️ Edit before importing
   * 🗑️ Delete if not suitable

<figure><img src="/files/hHS5WnWknDUqBNmXfRvq" alt=""><figcaption><p>Figure 4. Bot learns from messages panel</p></figcaption></figure>

<figure><img src="/files/bUB4Kgx6bq3OeqvIS0Tb" alt=""><figcaption><p>Figure 5. AI Actions button in Help Desk</p></figcaption></figure>

***


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://documents.cxgenie.ai/bots/bots-knowledge.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
