“We fixed that.”

Yes. In February. And May. There was also something in August, but we called that an improvement.

The customer has received two appointment reminders with different times.

Someone checks the booking, sends the correct time, and apologises. Sorted.

Until the next one.

The booking system sends a reminder. So does a separate service introduced when the booking system’s reminders weren’t very good.

They’re better now. Both are working beautifully.

Unfortunately, they disagree.

Someone adds a check before the reminders go out.

Now we have two reminder services and a person checking them.

At what point does the fix become another thing to look after?

There are good reasons for temporary solutions. A customer needs an answer today. The person who can change the system is away. You have permission to fix the immediate problem and absolutely no permission to touch the arrangement producing it.

So you do what you can.

The trouble comes when the temporary solution settles in. Nobody has time to revisit it because everyone is busy handling the exceptions.

You could give an AI assistant the checking job.

That might help immediately. It could flag conflicting times and prepare a correction for review.

But we should still ask why two services are sending the reminder.

Otherwise, we’ve hired a remarkably capable assistant to supervise an argument between two calendars.

I would like better things for everyone involved.

Start with one recurring failure. Find out where it begins.

In this case, someone changes an appointment. The booking system updates. The other service keeps the original time.

There’s the break.

Now we can ask a useful question: do we still need both services? If we do, how should a change reach them? Which system owns the appointment time?

Those answers give us something to fix.

Before changing anything, find out what the extra service still contributes. Perhaps it handles another kind of appointment. Perhaps it reaches customers the main system doesn’t.

Test the change with the people who manage bookings. Move an appointment. Cancel one. Check what the customer actually receives.

Then retire whatever is no longer needed, including the manual check.

Especially the manual check.

Otherwise, a conscientious person will keep doing it, because the last time someone said this was fixed, it wasn’t.

And write down what changed. Give someone responsibility for noticing if it fails again. The next person deserves more to work with than “ask Pat; there’s a thing.”

This doesn’t require rebuilding the business every time something breaks. Sometimes a small repair is exactly right.

But if the same repair keeps appearing on the list, its history is telling you something.

Your people may already know what that is. Ask what they keep correcting, what they’ve stopped reporting, and which “quick job” has become a permanent part of the week.

There’s a useful place to begin.

Before we fix it again.