On Monday 25 September 2017, at 07:28, my phone rang. It was the operations director of a client we had supported for years, and he was not calling about an error message.
07:28. The call.
Thieves had been through the premises over the weekend. The accounting server was gone. The payroll server was gone. Thirty desktop machines were gone. They had taken two large steel safes off the floor. When a crew removes safes, they are not leaving your servers behind.
This client ran 20 accounting databases and 10 payroll databases. Their payroll department was four people, and those four people had 2,500 employees to pay in three days. In South Africa, a missed payday at that scale is not an apology email. It is the difference between an awkward week and people who cannot buy food, and it can tip into large-scale industrial action.
Sit with the arithmetic for a moment. Twenty companies of processing history. Months of captured invoices, journals, payslips and reconciliations. Without offsite backups, the options that morning are recapturing months of work from paper, or closing.
08:13. On site.
We arrived 45 minutes after the call with every laptop we could pool, plus chargers and cables and very little conversation. The backups for this client were automated and offsite, which meant the thieves had taken hardware and left the history untouched. Restores began immediately, database by database, onto borrowed machines.
10:30. Payroll processing.
By 10:30 that morning the payroll team was working again. Every database restored. Three hours from the call. The replacement hardware only arrived two days later, which means the business was operating for two full days on machines it did not own, running data that had been stolen 36 hours earlier.
Every one of those 2,500 employees was paid on time.
The days after.
About a week later, the client's IT provider had restored roughly 95% of their general data from full-system backups. That deserves credit, and we gave it. Full-system recovery is their job and they did it well. But payroll could not wait a week, and it did not have to, because financial data lives in small, precise backup sets that restore in hours rather than days. The two services are not rivals. They are different answers to different clocks.
Why this story is on our website.
Because of the other calls. In the years since, I have answered the phone to a different Monday more than once: a crashed server, a ransomware note, a stolen machine, and the question that follows. Do you still have a backup from when you did my rollover three months ago?
When backups were a side task inside software support, the honest answer was sometimes no, because nobody had asked for them to be automated. Those businesses suffered, and there was nothing to restore. That answer is the reason Knox Findata exists as a service in its own right.
When you call, the answer must be yes. That is the entire product.