
How to Configure the Responding LiveAgent Tickets With an AI Agent-Spam Prevention Flow for Your Business
Learn how to customize API keys, URLs, and settings in the automated email and chat handling flow to match your business requirements.
Introduction
The Email & Chat Automation Flow comes pre-built with the core logic and structure already configured. However, to make it work for your specific business, you need to customize several key components: API credentials, URLs, knowledge base connections, and response policies. This guide walks you through exactly what needs to be changed and where to find each setting.
Required Changes Checklist
Before you start, gather these items:
- ✅ LiveAgent API key and workspace URL (if integrating)
- ✅ Your knowledge base documents uploaded to FlowHunt
- ✅ Your brand voice and support policy guidelines
Configuring API Credentials and URLs
Step 1: Replace LiveAgent API URLs
The flow contains placeholder URLs that need to be replaced with your actual LiveAgent workspace URL.
- Locate all API Request nodes in the flow. And click on the “Edit Template” button in the prompt connected to the URL section.
- Click each node to open settings
- Find the URL field containing “YOURLINK”
- Replace with your LiveAgent API endpoint (e.g.,
https://yourcompany.liveagent.com/api/v3/tickets
)

Step 2: Add Your LiveAgent API Key
Each API Request node needs authentication credentials.
- In the same API Request node settings
- Scroll to Query Parameters section
- Add or update the authorization header to the Create Data conected to this setion:
- apikey:
enter your V1 LiveAgent apikey
- apikey:

Connecting Your Knowledge Base
Step 1: Configure Document Retriever
The AI agent needs access to your support documentation.
- Click the Document Retriever node
- In Document Categories, select the categories containing your support docs
- Adjust Number of Results (default: 5-10 documents)
- Set Relevance Threshold (0.7 recommended)

Step 2: Verify Document Access
Before going live, ensure your documents are properly set up:
- Documents are uploaded to the Documents tab
- Categories are correctly assigned
- Content is current and accurate
Customizing Response Behavior
Step 1: Customize AI Agent System Message
This is the most important customization—it defines how your AI agent behaves.
- Click the Tool Calling Agent node
- Find the System Message field (large text area)
- Replace or modify the default guidelines

Key sections to customize:
Response Language:
Always respond in [YOUR LANGUAGE] regardless of input language.
Brand Voice:
Use a [formal/casual/friendly] tone. Address customers as [you/valued customer/etc].
Citation Rules:
When referencing documentation, cite sources as: [your format]
Include URLs in the format: [your format]
Limitations:
Do not discuss: [topics outside your scope]
Do not provide: [financial advice/medical info/etc]
Escalation Policy:
If unable to help, guide customers to: [contact method/ticket system]
Common Configuration Issues
API Connection Failed:
- Double-check URL format (include https://)
- Verify API key is correct and active
- Check workspace URL matches your LiveAgent account
Agent Not Finding Documents:
- Verify documents are in selected categories
- Check Document Retriever is connected to agent
- Ensure documents contain relevant content
Responses Not Following Guidelines:
- Review system message for clarity
- Check temperature isn’t too high (>0.7)
- Test with specific examples to refine prompts
Spam Detection Too Aggressive/Lenient:
- Adjust the spam classification prompt in OpenAI node
- Provide more specific examples of spam vs. legitimate messages
- Consider adding conditional logic for edge cases
Multi-Language Support
For international customers:
- Modify system message to detect input language
- Add language detection conditional logic
- Configure agent to respond in detected language
Going Live Checklist
Before deploying to production:
- ✅ All “YOURLINK” placeholders replaced
- ✅ API keys configured and tested
- ✅ Knowledge base documents verified
- ✅ System message customized for your brand
- ✅ Flow tested with realistic customer messages
By completing these configuration steps, your email and chat automation flow will be customized for your business and ready to handle customer interactions automatically.
Frequently asked questions
- What information do I need before configuring this flow?
You need your OpenAI API key, Anthropic API key, LiveAgent API credentials (if using), your LiveAgent workspace URL, and access to your knowledge base documents in FlowHunt.
- Where do I find the placeholder values that need to be replaced?
Look for 'YOURLINK' in API Request nodes, placeholder API keys in headers sections, and default system messages in the Tool Calling Agent. These all need to be customized for your setup.
- Do I need to configure both spam detection and the AI agent?
Yes, both components require configuration. The spam detection needs API setup (if using LiveAgent) and model credentials, while the AI agent needs your knowledge base connected and system message customized.
- Can I use this flow without LiveAgent?
Yes, you can remove or disable the LiveAgent API Request nodes if you're not using LiveAgent. The spam detection and AI agent will still function for direct chat interactions.