Challenge Tasks
Table of contents
- Challenge Tasks
Overview
Ready to level up? These optional challenges extend your email classification system with practical, real-world features that demonstrate your ability to integrate multiple services and build production-ready automations.
Challenge 1: Urgent Email Notifications
Objective
Send instant notifications via Telegram or Slack when urgent/high-priority emails arrive.
Requirements
- Detect urgent emails based on classification
- Send real-time alerts with email details
- Include sender, subject, and classification reason
- Support both Telegram and/or Slack
Why This Matters
Many urgent emails get buried in inboxes. Push notifications ensure critical messages get immediate attention, regardless of whether someone is checking email.
What You’ll Learn
- Integrating notification services (Telegram/Slack)
- Creating conditional notification logic
- Formatting rich notifications with email context
Challenge 2: Sentiment Trend Analysis
Objective
Track sentiment changes over time per sender to identify escalating customer situations.
Requirements
- Store historical sentiment data per sender
- Identify patterns indicating customer frustration
- Flag senders whose sentiment is declining
- Generate weekly trend reports
Why This Matters
A customer who sends three emails with increasingly negative sentiment is likely to churn. Early detection allows proactive intervention.
What You’ll Learn
- Time-series data analysis in Google Sheets
- Pattern recognition across multiple communications
- Building early warning systems for customer health
Challenge 3: Advanced Routing Rules
Objective
Implement complex, context-aware routing that considers multiple factors beyond just classification.
Requirements
- Time-based routing (business hours vs after-hours)
- VIP sender detection with special handling
- Keyword-triggered escalation
- Sender history consideration
Why This Matters
Real email systems need sophisticated routing. VIP customers expect faster responses, after-hours emails may need different handling, and certain keywords always trigger escalation.
What You’ll Learn
- Multi-factor decision logic
- Building lookup tables (VIP lists)
- Time-aware automation
- Escalation pathways
Challenge 4: Classification Confidence Tracking
Objective
Monitor AI classification confidence and route low-confidence emails for human review.
Requirements
- Extract confidence scores from AI responses
- Flag emails below confidence threshold (e.g., <70%)
- Create separate sheet/queue for human review
- Track accuracy by comparing AI vs human classification
Why This Matters
AI isn’t perfect. Production systems need to know when they’re uncertain and defer to humans. Tracking these cases improves the system over time.
What You’ll Learn
- Quality gates in AI systems
- Human-in-the-loop patterns
- Accuracy measurement and improvement
Challenge 5: Multi-Account Email Handling
Objective
Extend the system to monitor and classify emails from multiple Gmail accounts.
Requirements
- Connect 2+ Gmail accounts to n8n
- Classify emails from all accounts
- Track which account each email came from
- Maintain separate label/sheet tracking per account
Why This Matters
Businesses often manage multiple email accounts (support@, sales@, info@). A single automation should handle all of them efficiently.
What You’ll Learn
- Multi-account authentication
- Data isolation and tracking
- Scaling workflows across accounts
Saving Your Work
After completing any challenge, save your workflow to GitHub:
- Export Workflow: In n8n, click ⋯ menu → Download → Save as JSON
- Commit to GitHub: Follow the instructions in Part B to push your workflow
- Document Changes: Add a README explaining what you built
Getting Help
Stuck on a challenge?
- Check n8n documentation for specific nodes
- Review Part B to see how similar patterns work
- Search n8n community forums for similar use cases
- Remember: The goal is to learn by trying, not to get it perfect