Fallback workflow
A fallback workflow is the pre-planned alternative path an automated process follows when a step fails, times out or cannot reach a confident result.
What it means
Automation breaks: a bank portal is down, a scan is unreadable, a model returns nonsense. A fallback workflow decides in advance what happens next, instead of letting work vanish. It may be a retry, a switch to a simpler rule, a manual queue for staff, or a message to the owner. The fallback should keep money and stock records correct even if it is slower.
Example
A bakery chain's order reader cannot parse a handwritten indent from a hotel. The workflow sends the photo to the dispatch clerk's queue with a two-hour deadline, so the next morning's delivery is not missed.
Records and checks
- Record each time the fallback ran, why it started and who finished the work.
- Verify that no item remained in a failed state without an owner.
- The process owner reviews fallback frequency, since many runs signal a weak automated step.
Common mistakes
- Having no fallback, so failed items silently disappear.
- Making the manual route so slow that staff bypass it with side records.
Did not find your answer?Message the lamp team on WhatsApp. Ask about setup, your records, or anything we have not covered here.Ask us on WhatsApp →