# Knowledge

The **Knowledge** section allows you to provide data to your AI Agent.

This data is used by the agent to understand, retrieve, and respond to user queries using **Retrieval-Augmented Generation (RAG)**.

Without knowledge, the agent relies only on its general model.\
With knowledge, it becomes **context-aware and domain-specific**.

***

### How Knowledge Works

When a user asks a question:

1. The AI Agent searches the available knowledge sources
2. Relevant information is retrieved
3. The AI uses this data to generate a response
4. The response is grounded in your provided content

This ensures responses are:

* Accurate
* Contextual
* Based on your organization’s data

***

### Add Knowledge

You can add knowledge to your AI Agent in four different ways:

***

#### 1. Upload Files

Upload documents directly to the agent.

**Supported formats:**

* PDF (.pdf)
* Word (.doc, .docx)
* Text (.txt)
* CSV (.csv)
* PowerPoint (.pptx)
* JSON (.json)
* Markdown (.md)

**Limits:**

* Maximum **10 files per upload**
* Maximum **5 MB per file**

Use this for:

* Internal documentation
* Guides
* Policies
* Product documentation

***

#### 2. Public Websites

Add website URLs to be crawled and indexed.

The system will:

* Crawl the provided pages
* Extract content
* Make it available to the agent

Use this for:

* Documentation sites
* Help centers
* Public knowledge bases

***

#### 3. Applications

Connect external systems as knowledge sources.

Examples:

* ServiceNow Knowledge Base
* Xurrent Knowledge Articles

This allows the agent to:

* Access live data
* Use up-to-date knowledge
* Integrate directly with operational systems

***

#### 4. Text Input

Manually enter text as knowledge.

Use this for:

* Small pieces of information
* Instructions or notes
* Quick additions without uploading files

***

### Managing Knowledge

All added knowledge sources are listed in the **Knowledge** section.

Each item shows its status:

* **Ready** → Available for use
* **Processing** → Being indexed

You can also:

* Remove knowledge sources
* Update content (by re-uploading or editing)

***

### Best Practices

* Keep content **clean and structured**
* Avoid duplicate or conflicting information
* Use clear and descriptive language
* Regularly update knowledge sources
* Prioritize high-quality, relevant content

***

### When to Use Each Option

| Method          | Best For                          |
| --------------- | --------------------------------- |
| Upload Files    | Internal docs, structured content |
| Public Websites | External documentation            |
| Applications    | Live system data (ITSM, KBs)      |
| Text Input      | Quick or small additions          |

***

### What’s Next

After adding knowledge, your AI Agent can:

* Answer questions using your data
* Provide accurate and contextual responses
* Support users based on real documentation

You can further enhance your agent by:

* Adding [**Skills**](https://docs.autommate.com/userguide/ai-agent-composer/skills) (to take actions)
* Connecting [**Channels**](https://docs.autommate.com/userguide/ai-agent-composer/channels) (WhatsApp, Teams, Webhook, Mate Chat)

***

### Summary

Knowledge is the **context layer** of your AI Agent.

* It defines what the agent *knows*
* It improves response accuracy
* It enables domain-specific intelligence