Retry & Fallback Strategies¶
Definition¶
How to handle failures gracefully in eKYC — from individual step retries to complete fallback paths — minimizing user drop-off while maintaining security.
Retry Strategy per Step¶
| Step | Max Retries | Guidance on Retry | Fallback |
|---|---|---|---|
| Document capture | 3 | Specific quality feedback | Manual upload option |
| Selfie capture | 3 | Lighting/position guidance | Active liveness as fallback |
| Face match fail | 1 retry with re-capture | "Ensure good lighting, remove glasses" | Manual review |
| Liveness fail | 2 | "Hold still, ensure good lighting" | V-KYC escalation |
| OCR low confidence | 1 re-capture | "Ensure no glare on document" | Manual data entry |
| Database timeout | 3 auto-retries | Invisible to user | Queue for async retry |
| Complete failure | — | — | V-KYC or branch visit |
Fallback Chain¶
graph TD
A[Automated eKYC] -->|Fails 3x| B[Simplified eKYC<br/>Relaxed thresholds + manual review]
B -->|Still fails| C[Video KYC<br/>Human agent assistance]
C -->|Not available| D[Branch Visit<br/>In-person verification]Key Takeaways¶
Summary
- Specific guidance on retry ("Move into better light") dramatically improves retry success
- Fallback chain ensures every user has a path to completion
- V-KYC is the universal fallback when automated eKYC can't resolve
- Invisible retries (database timeouts) should be handled automatically