n8n Workflow Repair & Optimization
Budget: ₹600 – ₹1,500 INR
Project Overview
Fix critical errors and optimize an existing n8n workflow for a WhatsApp-based customer service automation system (Mariam CS). The workflow handles tenant/landlord intent detection, lead qualification, and property search assistance.
Deliverables
1. Critical Error Fixes (MUST FIX)
Estimated Time: 3-4 hours
Fix Missing Trigger References (6 nodes affected)
Update all nodes referencing non-existent "WhatsApp Trigger" to use "WhatsApp Trigger2"
Nodes: check if lead already been processed, Create a record2, Send message and wait for response, Search records in Airtable1, Simple Memory2, Search records in Airtable2, If
Reconnect Disconnected Workflow Branch
Connect If2 node to WhatsApp Trigger2
Verify entire audio transcription branch is properly connected
Fix AI Agent Configurations
Connect language model to Intent Checker node (currently missing)
Configure system messages for all AI Agent nodes (4 nodes)
Add dynamic expressions to AI Agent text fields where missing
Fix Data Reference Issues
Correct Search records node filter formula (invalid field reference)
2. Best Practice Improvements (RECOMMENDED)
Estimated Time: 2-3 hours
AI Agent Optimization
Separate system messages from user prompts in all AI Agent nodes
Add proper system message configuration to: AI Agent2, AI Agent3, Intent Checker, AI Agent
Implement dynamic expressions in agent text fields
Tool Node Configuration
Add dynamic expressions to AI Agent Tool3 parameters
Ensure proper $fromAI() usage where applicable
Workflow Logic Review
Document and clarify the purpose of disabled nodes
Recommend removal or activation of incomplete features
3. Testing & Validation (REQUIRED)
Estimated Time: 2 hours
End-to-End Testing
Test WhatsApp message reception (text messages)
Test audio message transcription flow
Test intent detection (tenant vs landlord)
Test lead qualification flow
Test property search assistant flow
Error Handling Verification
Verify all error paths work correctly
Test edge cases (empty responses, API failures)
Documentation
Provide workflow execution report
Document any remaining issues or limitations
Technical Requirements
Skills Required
n8n Platform: Advanced knowledge of n8n workflow automation
AI/LLM Integration: Experience with OpenAI, Gemini, and LangChain nodes
API Integration: WhatsApp Business API, Airtable API
JavaScript: For Code node debugging and optimization
Workflow Logic: Understanding of conditional routing and data transformation
Access Needed
n8n workflow editor access (read/write)
Workflow JSON export/import capability
Test WhatsApp account for validation
Documentation of expected workflow behavior
Timeline & Milestones
Phase 1: Critical Fixes (Day 1-2)
Fix all trigger references
Reconnect disconnected branches
Fix AI Agent language model connections
Deliverable: Workflow executes without critical errors
Phase 2: Optimization (Day 3)
Configure AI Agent system messages
Add dynamic expressions
Fix data reference issues
Deliverable: Workflow follows best practices
Phase 3: Testing & Handover (Day 4)
Complete end-to-end testing
Document findings and recommendations
Deliverable: Test report + optimized workflow
What's NOT Included
New feature development
AWS S3 configuration (disabled nodes)
Media upload functionality (disabled nodes)
Changes to external systems (Airtable schema, WhatsApp setup)
Ongoing maintenance or support beyond delivery
Success Criteria
- Workflow executes without errors
- All trigger references are correct
- AI Agents have proper language model connections
- System messages properly configured
- Test execution completes successfully for all paths
- Documentation provided for any remaining limitations
Questions for Client
Are the disabled nodes (AWS S3, media upload) needed? Should they be removed or completed?
Do you have test credentials for WhatsApp and Airtable?
What is the expected behavior for the disconnected If2 branch?
Are there specific test scenarios you want validated?
Do you need training on maintaining the workflow after fixes?
Next Steps
Award Project: Accept proposal and provide access
Kickoff Call: 30-minute call to clarify requirements (optional)
Work Begins: Start with Phase 1 critical fixes
Daily Updates: Progress reports via Freelancer.com messaging
Final Delivery: Tested workflow + documentation
Ready to fix your workflow and get it running smoothly!
Fix critical errors and optimize an existing n8n workflow for a WhatsApp-based customer service automation system (Mariam CS). The workflow handles tenant/landlord intent detection, lead qualification, and property search assistance.
Deliverables
1. Critical Error Fixes (MUST FIX)
Estimated Time: 3-4 hours
Fix Missing Trigger References (6 nodes affected)
Update all nodes referencing non-existent "WhatsApp Trigger" to use "WhatsApp Trigger2"
Nodes: check if lead already been processed, Create a record2, Send message and wait for response, Search records in Airtable1, Simple Memory2, Search records in Airtable2, If
Reconnect Disconnected Workflow Branch
Connect If2 node to WhatsApp Trigger2
Verify entire audio transcription branch is properly connected
Fix AI Agent Configurations
Connect language model to Intent Checker node (currently missing)
Configure system messages for all AI Agent nodes (4 nodes)
Add dynamic expressions to AI Agent text fields where missing
Fix Data Reference Issues
Correct Search records node filter formula (invalid field reference)
2. Best Practice Improvements (RECOMMENDED)
Estimated Time: 2-3 hours
AI Agent Optimization
Separate system messages from user prompts in all AI Agent nodes
Add proper system message configuration to: AI Agent2, AI Agent3, Intent Checker, AI Agent
Implement dynamic expressions in agent text fields
Tool Node Configuration
Add dynamic expressions to AI Agent Tool3 parameters
Ensure proper $fromAI() usage where applicable
Workflow Logic Review
Document and clarify the purpose of disabled nodes
Recommend removal or activation of incomplete features
3. Testing & Validation (REQUIRED)
Estimated Time: 2 hours
End-to-End Testing
Test WhatsApp message reception (text messages)
Test audio message transcription flow
Test intent detection (tenant vs landlord)
Test lead qualification flow
Test property search assistant flow
Error Handling Verification
Verify all error paths work correctly
Test edge cases (empty responses, API failures)
Documentation
Provide workflow execution report
Document any remaining issues or limitations
Technical Requirements
Skills Required
n8n Platform: Advanced knowledge of n8n workflow automation
AI/LLM Integration: Experience with OpenAI, Gemini, and LangChain nodes
API Integration: WhatsApp Business API, Airtable API
JavaScript: For Code node debugging and optimization
Workflow Logic: Understanding of conditional routing and data transformation
Access Needed
n8n workflow editor access (read/write)
Workflow JSON export/import capability
Test WhatsApp account for validation
Documentation of expected workflow behavior
Timeline & Milestones
Phase 1: Critical Fixes (Day 1-2)
Fix all trigger references
Reconnect disconnected branches
Fix AI Agent language model connections
Deliverable: Workflow executes without critical errors
Phase 2: Optimization (Day 3)
Configure AI Agent system messages
Add dynamic expressions
Fix data reference issues
Deliverable: Workflow follows best practices
Phase 3: Testing & Handover (Day 4)
Complete end-to-end testing
Document findings and recommendations
Deliverable: Test report + optimized workflow
What's NOT Included
New feature development
AWS S3 configuration (disabled nodes)
Media upload functionality (disabled nodes)
Changes to external systems (Airtable schema, WhatsApp setup)
Ongoing maintenance or support beyond delivery
Success Criteria
- Workflow executes without errors
- All trigger references are correct
- AI Agents have proper language model connections
- System messages properly configured
- Test execution completes successfully for all paths
- Documentation provided for any remaining limitations
Questions for Client
Are the disabled nodes (AWS S3, media upload) needed? Should they be removed or completed?
Do you have test credentials for WhatsApp and Airtable?
What is the expected behavior for the disconnected If2 branch?
Are there specific test scenarios you want validated?
Do you need training on maintaining the workflow after fixes?
Next Steps
Award Project: Accept proposal and provide access
Kickoff Call: 30-minute call to clarify requirements (optional)
Work Begins: Start with Phase 1 critical fixes
Daily Updates: Progress reports via Freelancer.com messaging
Final Delivery: Tested workflow + documentation
Ready to fix your workflow and get it running smoothly!