Table of Contents

AI Agents Overview

Pau Sanchez Updated by Pau Sanchez

Overview

AI agents provide intelligent conversational capabilities that integrate seamlessly with your existing chatbot flows. This guide covers the setup process, core components, and practical implementation based on a lead generation example for an English Academy.

Core Components

Knowledge Base (Left Side Panel)

The knowledge base contains all the specific information your agent will reference during conversations.

What to Include:

  • Company or service-specific knowledge
  • Common FAQs and their answers
  • Any detailed information the agent should know about your business

How It Works: The agent uses this knowledge base when users ask specific questions or need information about your services. All uploaded content becomes part of the agent's understanding and response capability.

Agent Instructions (Center Panel)

This is the core of your AI agent configuration. The instructions define exactly how your agent should behave and operate.

Key Configuration Areas:

  • Behavior Guidelines: How the agent should interact with users
  • Response Style: How the agent should answer questions
  • Edge Case Handling: Instructions for managing unusual situations
  • Information Collection: Specific guidance on what data to gather and how

Important Connection: The instructions you write here directly relate to the custom information fields you define. The agent will follow these instructions to collect the specific data you need.

Information Capture (Right Side Panel)

Configure what information your agent should collect from users during conversations.

Standard Fields Available:

  • Name
  • Email
  • Company
  • Phone

Custom Fields: You can add specific information fields relevant to your business needs. In the English Academy example, custom fields included:

  • Course interest (business courses, etc.)
  • Budget range (under 22 euros, etc.)

Critical Setup Tip: When creating custom fields, provide specific examples and instructions in the Agent Instructions section. This ensures the agent captures information exactly as you need it.

Exit Conditions (Right Side Panel)

AI agents operate in a conversational loop. Exit conditions tell the agent when to stop the conversation and return control to your static chatbot flow.

How It Works:

  • The agent continues conversing until specific conditions are met
  • Once conditions are satisfied, the conversation transfers to a designated block in your static chatbot
  • This allows integration with APIs, CRM systems, or continued automated flows

Implementation Workflow

Setting Up the Complete Flow

The implementation involves three main components working together:

  1. Static Chatbot Flow: Your existing chatbot structure
  2. AI Agent: The intelligent conversational component
  3. Integration Points: Where the agent connects back to static flow
Practical Example: English Academy Lead Generation

Flow Structure:

  1. Initial Static Flow: Basic chatbot asking for name
  2. Jump to AI Agent: "Jump to Block" transfers control to the AI agent
  3. Agent Conversation: AI takes over and handles intelligent conversation
  4. Information Collection: Agent gathers course interest and budget information
  5. Exit and Return: Agent returns control to static flow for CRM integration
Block ID Integration

Visual Connection Method:

  • Each block in your chatbot has a unique Block ID

  • Reference specific Block IDs in your exit conditions
  • This creates a visual connection showing exactly where conversations will continue
  • Makes the flow easier to understand and less likely to break

Example from Demo: The exit condition referenced a specific block ID that was visually connected in the flow, making it clear where users would be directed after the AI agent completes its tasks.

Live Implementation Example

The Conversation Flow

Welcome Message: The agent starts with a predefined welcome message, then begins gathering the requested information.

Information Gathering Process: In the English Academy example, the agent was instructed to collect two pieces of information:

  1. Course interest
  2. Budget range

User Interaction Example:

  • User responds: "Business"
  • Agent captures: "Business English" (matches the intended category)
  • User responds: "22"
  • Agent processes: Categorizes into appropriate budget range as defined in instructions (0-100)

Intelligent Processing: The agent demonstrates flexibility in understanding user responses. When a user says simply "business," the agent correctly interprets this as "Business English" based on the context and instructions.

Data Storage and Integration

Automatic Information Storage: Once the agent collects the required information, it automatically stores the data based on the instructions provided.

Budget Range Processing: The demo showed how budget responses are intelligently categorized into predefined ranges rather than storing exact amounts. This was specifically configured in the agent instructions.

Field Accessibility: The collected information becomes available for use in multiple contexts:

  • Within the AI agent conversation
  • In the static chatbot flow after handoff
  • For API integrations and CRM systems
  • For display back to users
Exit Process

Automatic Exit Trigger: In the example, the agent was configured to exit immediately after gathering all required information (course interest and budget).

Seamless Handoff: Once the exit condition was met, the conversation seamlessly returned to the static chatbot flow, where the collected information was immediately available for further processing or display.

Configuration Best Practices

Writing Effective Instructions

Be Specific About Data Collection: When defining what information to collect, include specific examples in your instructions. The English Academy example showed how providing examples of budget ranges and course types ensures accurate data capture.

Handle Variations: Account for different ways users might express the same information. In the demo, users could say "business" instead of "business course" and the agent would still capture the correct information.

Define Clear Categories: For fields like budget ranges, define specific categories in your instructions rather than collecting exact amounts. This makes the data more useful for later processing and routing.

Knowledge Base Optimization

Upload Relevant Content: Include comprehensive information about your business that the agent might need to reference during conversations.

Keep Information Current: Regularly update your knowledge base to ensure the agent has access to the most current information about your services.

Testing Your Setup

Verify Data Collection: Test that the agent captures information exactly as specified in your instructions.

Check Exit Conditions: Ensure the agent exits at the right time and transfers to the correct block in your static flow.

Test Integration: Verify that collected information is properly accessible in your static chatbot flow and any connected systems.

Integration with Existing Systems after Exit condition

Once the AI agent completes its task and hands control back to your static chatbot flow, all the collected information becomes available for integration with external systems. Here are five common use cases for leveraging the gathered data:

CRM Lead Creation

Automatically create new leads in your CRM with all collected information (name, email, course interest, budget) and assign to appropriate sales team based on qualification criteria.

Email Marketing Automation

Trigger personalized email sequences based on collected preferences - send business course information to users who expressed interest in business English, with content tailored to their budget range.

Calendar Integration

For qualified leads meeting specific criteria (e.g., premium budget range), automatically send calendar booking links or schedule follow-up calls with sales representatives.

Customer Database Updates

Update existing customer profiles with new information or create new entries, ensuring all collected data is stored for future reference and segmentation.

Real-Time Notifications

Send instant alerts to sales teams when high-value leads are identified, enabling immediate follow-up while the prospect's interest is highest.

Troubleshooting Common Issues

Information Not Collecting Properly

Check Your Instructions: Ensure your agent instructions include specific examples of how information should be collected and categorized.

Verify Field Configuration: Make sure custom fields are properly defined and connected to your instructions.

Exit Conditions Not Working

Verify Block IDs: Double-check that referenced Block IDs exist and are correctly specified in your exit conditions.

Review Condition Logic: Ensure exit conditions are clearly defined and achievable based on your information collection setup.

Agent Not Using Knowledge Base

Content Format: Verify that uploaded content is properly formatted and comprehensive enough for the agent to reference.

Instruction Clarity: Make sure your instructions guide the agent on when and how to use knowledge base information.

Conclusion

AI agents provide a powerful way to enhance your chatbot with intelligent conversation capabilities while maintaining seamless integration with your existing flows. The key to successful implementation lies in:

  1. Clear Instructions: Detailed guidance on agent behavior and data collection
  2. Proper Field Configuration: Well-defined custom fields with examples
  3. Effective Exit Conditions: Clear criteria for returning to static flows
  4. Comprehensive Knowledge Base: Relevant, up-to-date information for agent reference

By following the patterns demonstrated in the English Academy example, you can create sophisticated lead generation and customer interaction flows that gather valuable information while providing personalized user experiences.

Was this article helpful?

AI Lead Gen Assistant

Contact