> For the complete documentation index, see [llms.txt](https://angoor-ai.gitbook.io/angoor-ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://angoor-ai.gitbook.io/angoor-ai/basics/configuration/agents.md).

# Agents

The Agents section enables you to create AI-powered conversational assistants that automate customer interactions, answer inquiries, and execute specific tasks tailored to your business needs.

## Agents Dashboard

Your central hub for managing all AI agents displays comprehensive configuration and performance metrics:

* **Agent Name:** Unique identifier for your AI assistant (e.g., "Customer Support Bot", "Lead Qualifier")
* **Creation Details:** Timestamp and owner information for tracking and accountability
* **Variables:** Count of dynamic placeholders for personalized interactions
* **Knowledge Base:** Connected information sources powering accurate responses
* **Description:** Purpose and capabilities summary for team reference
* **Tags:** Organizational labels for filtering and categorization

<figure><img src="/files/GA26TkZWqcISURjKUifd" alt=""><figcaption></figcaption></figure>

## Creating Your First Agent

**Step 1:** Click "Create Agent" and enter a descriptive name that clearly indicates the agent's purpose.

**Step 2:** Define your agent's personality and behavior in the Core tab using structured prompts and instructions.

**Step 3:** Configure technical settings including AI model selection, response parameters, and custom features.

**Step 4:** Connect knowledge bases and forms to provide context and enable data collection capabilities.

**Step 5**: Test your agent with sample conversations before deploying to production.

<figure><img src="/files/Io1kPs0954zS5tLQacev" alt=""><figcaption></figcaption></figure>

## Core Tabs

The Core tab is your workspace for crafting agent personality, knowledge integration, and conversation logic.

<figure><img src="/files/gY3yu0WmK7mxMg7MI5UI" alt=""><figcaption></figcaption></figure>

### **Prompt Engineering**&#x20;

Create comprehensive instructions using a structured template approach:

```
## Job Description
Define what your agent does and its primary responsibilities

## Character
Establish personality traits and communication style

## Information to Collect
Specify data points to gather during conversations

## Tone and Personality
Set the emotional intelligence and interaction approach

## Style and Language
Define vocabulary, formality level, and cultural considerations
```

For comprehensive guidance on advanced prompting techniques and best practices, refer to these detailed documentation resources: [OpenAI Prompting Guide](https://cookbook.openai.com/examples/gpt4-1_prompting_guide) & [Anthropic Claude 4 Best Practices](https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/claude-4-best-practices).

### **Dynamic Variables**&#x20;

Add personalization elements that adapt to each conversation:

* Click "Add Variable" to create placeholders for user data
* Reference variables in prompts using {{ }}, {{ }} format or by reference variable button.
* You populate these variables at any point in the flow

### **Knowledge Reference**&#x20;

Enhance agent intelligence with organizational information:

* Use "Refer Knowledge Base" to connect documentation, FAQs, and product information
* Reference specific knowledge sections within prompts for contextual responses
* Combine multiple knowledge sources for comprehensive coverage

### **Conditional Logic**&#x20;

&#x20;Create intelligent response paths with flexible conditional logic:

* Add conditions using Field-Operator-Value structure
* Design different conversation flows for various scenarios
* Stack multiple conditions for complex decision trees

## Settings Tabs

<figure><img src="/files/zss3BZBYc0VPBAFVPhoW" alt=""><figcaption></figcaption></figure>

#### Metadata

Configure tool name, description, and tags for clear identification and organization across your workspace.

#### Model Selection&#x20;

Choose from GPT 4.1, GPT 3.5, or custom models with token capacities of 8K, 16K, 32K, or 100K and Standard or Fast response times. More models will be available soon