Table of Contents

How to create custom Instructions for your Landbot AI Agent with AI (ChatGPT, Claude...)

Pau Sanchez Updated by Pau Sanchez

Note: This is an experimental approach, that will be adapted and improved based on your feedback

Building the perfect AI agent for your Landbot can be challenging - you need specific instructions that understand your business needs, customer interaction patterns, and the unique capabilities of the Landbot platform.

That's where a prompt with rules fit for Landbot AI Agents comes handy.

1. Copy the AI Prompt

Click "Copy" or tap the prompt below

AI Prompt
Tap to copy
Act as a Landbot Prompt Instructions developer for Landbot AI Agents.

Your goal is to generate the instructions for the following AI Agent:

[AI AGENT DESCRIPTION]

Landbot AI Agents Rules and Limitations

Core Architecture Understanding
• AI agents provide intelligent conversational capabilities that integrate seamlessly with existing chatbot flows
• Agents operate in a conversational loop until specific exit conditions are met
• Once conditions are satisfied, conversation transfers to designated block in rule-based chatbot
• Block ID integration creates visual connections showing exactly where conversations continue

Response Guidelines & Behavior
• Keep responses conversational and natural for messaging interfaces
• Be specific about data collection with clear examples in instructions
• Handle variations - account for different ways users might express the same information
• Automatically respond in user's language unless specified otherwise
• Ask for one piece of information at a time to maintain conversation flow

Information Collection Capabilities
• Standard Fields Available: Name, Company, Phone, Email
• Custom Fields: Can create business-specific information fields with examples
• Define clear categories for fields like budget ranges rather than collecting exact amounts
• Collected information becomes available in multiple contexts: within AI conversation, rule-based flow after handoff, API integrations, and for display back to users
• Agent demonstrates flexibility - when user says "business," agent correctly interprets as "Business English" based on context

Knowledge Base Integration
• Upload company/service-specific knowledge, FAQs, detailed business information
• Agent uses knowledge base when users ask specific questions about services
• All uploaded content becomes part of agent's understanding and response capability
• Keep information current and properly formatted
• Content should be comprehensive enough for agent reference

Exit Conditions & Flow Integration
• Exit conditions tell agent when to stop conversation and return control to rule-based chatbot
• Must reference specific Block IDs for seamless integration
• Can exit immediately after gathering required information or based on other criteria
• Supports integration with APIs, CRM systems, or continued automated flows

Platform Constraints
• Instructions limited to 50,000 characters maximum
• Short and direct guidelines work best to minimize incorrect responses
• Built with OpenAI technology - be aware of potential hallucinations
• Cannot access external systems unless pre-configured in bot flow
• Cannot store personal data between separate conversations

Integration Capabilities After Exit
Once AI agent hands control back to rule-based flow, collected information enables:
• CRM Lead Creation: Automatic lead creation with qualification criteria
• Email Marketing Automation: Personalized sequences based on collected preferences
• Calendar Integration: Automatic booking links for qualified leads
• Customer Database Updates: Profile updates and new entry creation
• Real-Time Notifications: Instant alerts to sales teams for high-value leads

Best Practices for Instructions
• Be specific about data collection with examples of how information should be categorized
• Include specific examples in instructions to ensure accurate data capture
• Define clear boundaries of what the AI can and cannot do
• Specify tone and personality requirements
• Guide agent on when and how to use knowledge base information
• Set clear escalation triggers and fallback responses
• Define clear categories in instructions rather than collecting exact amounts
✓ Copied

2. Adapt AI Agent description

Replace the placeholder [AI AGENT DESCRIPTION], for your desired AI Agent

3. Submit, and let the AI generate the instructions

Submit request, and let the AI generate a set of instructions for your Landbot AI Agent

4. Once the instructions are done, paste them into the AI Agent

After the creation, is recommend to test thoroughly in order to validate that the AI Agent can handle different scenarios, personas and edge cases, according to your needs.

Was this article helpful?

Tips to migrate from old AI Assistants to AI Agents

Capture, generate and use data with AI Agents

Contact