> ## Documentation Index
> Fetch the complete documentation index at: https://koreai.mintlify.app/llms.txt
> Use this file to discover all available pages before exploring further.
# Empower Agents with Knowledge
Enable context-aware responses with RAG-powered retrieval.
## Overview
Knowledge connects your agents to data sources through **Knowledge Tools** — intelligent components that use Search AI's Retrieval-Augmented Generation (RAG) capabilities. Instead of relying solely on the LLM's training data, agents retrieve relevant information from your knowledge bases and incorporate it into responses. Integrated through the **AI for Service** platform, Knowledge Tools enable agents to understand user intent, retrieve relevant content, and generate accurate, business-aware responses in real time.
```
User Question → Semantic Search → Retrieve Relevant Chunks → Augment Prompt → Generate Response
```
## How It Works
The RAG Pipeline:
```mermaid actions={false} theme={null}
flowchart LR
A["**1. User Query**
'What's your return policy for electronics?'"]
B["**2. Query Embedding**
Convert question to vector representation"]
C["**3. Semantic Search**
Search knowledge base for similar content"]
D["**4. Chunk Retrieval**
Return top-k relevant passages from documents"]
E["**5. Prompt Augmentation**
Inject retrieved chunks as context into agent's prompt"]
F["**6. Response Generation**
Agent generates grounded, accurate response"]
A --> B --> C
D --> E --> F
```
## Key Capabilities
| Capability | Details |
| ------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Extensive Connectors** | Connect to 100+ data sources:
Cloud Storage (Google Drive, OneDrive, Dropbox),
Collaboration (Confluence, SharePoint, Notion),
Business Apps (Salesforce, Jira, HubSpot),
Communication (Slack, Microsoft Teams),
Custom Apps |
| **Intelligent Search** | Hybrid search (keyword + semantic matching),
Multi-vector (weighted embedding strategies),
Programmable extraction (custom chunking pipelines) |
| **Enterprise Security** | Access control integration, permission-aware retrieval, audit logging |
## Setting Up Knowledge
The **Knowledge** page of your Agentic App provides a structured view of all connected knowledge bases, including their total count, associated agents, and management options from a central location.
### Step 1: Create a Knowledge Tool
Navigate to **Knowledge** in your app and click the **Create Knowledge Base** drop-down. Choose one of the following:
* **Create from existing sources** — Link the Agentic App to an existing Search AI application in the same workspace.
* **Create from scratch** — Create a new Search AI application.
#### Create from Existing Sources
Configure the following details:
* **Knowledge Base Name** — Unique name to identify the knowledge tool in the Agentic App.
* **Description** — Brief summary of the content included. Helps agents determine when to use this knowledge base.
* **Knowledge Sources** — Click *Link Knowledge Source* to select an existing Search AI application from the same workspace.
After selecting the app, set the retrieval configuration:
* **Use Knowledge Base to respond with**:
* **Chunks** — Returns the top relevant passages from the knowledge base.
* **Answer** — Returns an answer generated by Search AI.
* Selecting both returns the generated answer along with the supporting chunks.
* **Meta Filters (Optional)** — Filter retrieved content by metadata. For example, to retrieve only web content:
```json theme={null}
[
{
"condition": "AND",
"rules": [
{
"fieldName": "sourceType",
"fieldValue": ["web"],
"operator": "contains"
}
]
}
]
```
Click **Create** to complete the integration.
#### Create from Scratch
Enter a **Knowledge Base Name** and **Description**, then click **Create**. This creates a new Search AI application in the same workspace. Next, navigate to Search AI to:
* Set extraction strategies
* Ingest content sources
* Set retrieval strategies
* Configure answer generation
### Step 2: Test the Knowledge Tool
Click **Test your knowledge base** and provide input parameters to validate the tool before deployment. The response includes output and detailed logs for troubleshooting, performance analysis, and refining query behavior.
### Step 3: Link to Agent
1. Open your agent configuration.
2. Navigate to **Knowledge**.
3. Select the knowledge tool.
4. Configure retrieval settings.
### Other Operations
* **Delete** — Remove a knowledge tool from the app via the Knowledge page. The underlying Search AI application remains available in the workspace and can be linked to other Agentic Apps.
* **Manage Content** — Use the **Manage Content** option to navigate to the Search AI app and update indexed content or configurations.
## Accessing and Managing Search AI Applications
Click **Search AI** to view and manage all Search AI applications in your workspace. These can be integrated into Agentic Apps as Knowledge Tools. The **Linked Apps** column shows which Agentic Apps use each integration.
To create a new Search AI application from this page:
1. Click **+ New Integration**.
2. Provide a name and description.
3. Click **Manage Content** to open the app in AI for Service and configure it.
To include an integration in your Agentic App, go to the app's Knowledge page and add it.
## Example Scenario
An agent designed to answer questions about a company's products can configure a Knowledge Tool that indexes the company's website using Search AI. When a user asks *"What AI solutions does the company offer?"*, the agent retrieves information directly from the indexed content and delivers an accurate, up-to-date answer.
```mermaid actions={false} theme={null}
flowchart LR
User(["Query ->
Response <-"])
Agent["AI Agent"]
KT["Knowledge Tool"]
SAI["Search AI"]
DB[("DB")]
Web["Web"]
Files["Files"]
User <--> Agent
Agent <--> KT
KT <--> SAI
SAI --> DB
SAI --> Web
SAI --> Files
```
## Answer Generation Examples
### Extractive
Pull exact quotes from source documents.
```
Question: "What's the return window?"
Retrieved: "Items may be returned within 30 days of purchase."
Answer: "Items may be returned within 30 days of purchase."
[Source: Return Policy, Section 2.1]
```
### Generative
Synthesize answers using retrieved context.
```
Question: "Can I return opened electronics?"
Retrieved:
- "Electronics have a 15-day return window."
- "Opened items may be subject to restocking fee."
- "Defective items can be returned within warranty period."
Answer: "Yes, you can return opened electronics within 15 days,
though a restocking fee may apply. If the item is
defective, you can return it within the warranty period
without a fee."
```
***
## Best Practices
### Quality Content
* Keep documentation up-to-date
* Use clear, structured writing
* Include relevant metadata
### Appropriate Chunking
| Content Type | Strategy | Chunk Size |
| ------------- | --------------------- | ----------------------- |
| FAQs | Question-answer pairs | Small (200-300 tokens) |
| Documentation | Sections/headers | Medium (400-600 tokens) |
| Long articles | Semantic breaks | Variable |
### Retrieval Tuning
* Start with top\_k=5, adjust based on results
* Use score thresholds to filter low-quality matches
* Enable reranking for better precision
### Grounding Instructions
Tell agents how to use knowledge:
```markdown theme={null}
Instructions:
- Base your answers on the retrieved knowledge
- If the answer isn't in the knowledge base, say so
- Cite sources when providing specific information
- Don't make up information not present in the context
```
[Learn More](/ai-for-service/searchai/tips-and-faqs).
## Monitoring
Track knowledge retrieval performance:
* **Hit rate**: How often relevant content is found
* **Latency**: Search response times
* **Relevance**: User satisfaction with answers
* **Coverage**: Questions without good matches
[Learn More](/ai-for-service/analytics/searchai/answer-insights).
***