A routine can seem reliable until something unusual happens. A document is missing, a supplier responds late, or an approval waits while a colleague is away.
Someone finds a workaround. The task is completed. A few weeks later, the same thing happens again.
The useful question is not how to prevent every exception. It is how to recognise the ones that deserve a lasting, modest change.
Notice what repeats
An isolated problem may need no new rule. A repeated interruption is different: it consumes attention each time, even when the immediate fix is easy.
Consider invoices that regularly arrive without a purchase-order reference. Each can be corrected manually. But repeated corrections suggest that the information needs to be gathered earlier.
The repetition, rather than any single invoice, tells the team where the process is weak.
Keep the record light
For a few weeks, note the date, the process, what interrupted it, the immediate workaround and who can decide whether to change something.
A shared list inside an existing tool is usually enough. Avoid lengthy reports or unnecessary personal details; the point is to understand a process, not to create another administrative task.
Separate today's repair from tomorrow's improvement
An immediate repair gets the current task moving. A process improvement makes the next occurrence less likely.
If approvals stall when the usual reviewer is away, today's repair may be finding an authorised substitute. The improvement is a clear backup arrangement for future absences.
Both actions matter, but they solve different problems.
Decide what deserves attention
Look at frequency, consequences and the effort required to prevent a repeat.
A rare formatting issue may be acceptable. A minor fault that delays customers every week may deserve attention. Where access, money or personal information is involved, even an infrequent exception can matter.
Do not turn every inconvenience into a new policy. Prefer the smallest change that addresses the underlying cause.
Test a small change
Sometimes a clearer instruction, a better template or an earlier prompt is enough. If requests often arrive missing one important detail, try asking for that detail at the start rather than adding a long form.
Then check whether the number of follow-ups actually falls. If it does not, adjust the approach instead of preserving a rule that creates work without helping.
Close the loop
Use a few minutes in an existing weekly check-in to decide which repeated exceptions need action and which can be accepted.
Record whether a change worked. The best measure is not the number of procedures written, but fewer interruptions and less time spent rescuing the same task.
Small teams do not need elaborate incident management for every inconvenience. They do benefit from noticing when a clever workaround is becoming a permanent burden.