The Problem
Errors frustrate users and often lead to abandonment:- Cryptic error messages they don’t understand
- No clear path to resolution
- Feeling stuck and helpless
- Lost work or progress
- Distrust in the platform
The Solution
Turn error moments into support opportunities with contextual AI assistance.Implementation
Step 1: Create Error Events
Event:error_payment_failed
error_api_failure
Step 2: Integrate with Error Handling
Step 3: Handle Specific Error Types
Payment ErrorsError Types and Responses
| Error Type | Trigger Style | Tone |
|---|---|---|
| Payment Failed | Toast (urgent) | Empathetic, solution-focused |
| Validation Error | Toast | Helpful, clarifying |
| Rate Limited | Banner | Informative, educational |
| Server Error | Toast | Apologetic, reassuring |
| Network Error | Toast | Technical, troubleshooting |
| Permission Denied | Toast | Explanatory, directive |
Best Practices
Be empathetic, not robotic
Be empathetic, not robotic
❌ “Error code 422: Validation failed”
✅ “Oops! A few things need fixing. Want me to help?”
Don't trigger on every error
Don't trigger on every error
Skip triggers for:
- 401 (unauthorized) - just redirect to login
- 404 (not found) - show 404 page
- Minor validation errors
Include actionable context
Include actionable context
Tell your agent what the user was trying to do:
- What action they attempted
- What went wrong
- What solutions are available
Log for debugging
Log for debugging
Always include request IDs and error codes so support
can investigate if the user escalates.
Measuring Success
| Metric | Description |
|---|---|
| Error Resolution Rate | % of errors resolved after chat |
| Abandonment After Error | Users who leave after an error |
| Support Escalation Rate | Errors requiring human support |
| Time to Resolution | How quickly users recover |