Nigerian Bank Moves ATM Operations to Predictive Failure Intelligence With Ollandi
A Nigerian commercial bank uses Ollandi for ATM to open cases before service thresholds are breached, to separate terminal faults from switch, power and carrier failures, and to treat missing evidence as a risk instead of a healthy terminal.
The condition
The bank's ATM team learned about most failures from customers. A terminal would stop dispensing while its monitoring status stayed green, and the first signal was a complaint at the branch or on social media.
Its estate also depends on conditions outside the terminal. Sites switch between mains supply, standby generators and UPS reserve through the day. Carrier links fail over. A degraded connection to the switch can take down a whole region of terminals that are physically fine. Each of these produced a flood of individual alerts, and engineers were often sent to terminals that had nothing wrong with them.
Ollandi's role
The bank deployed Ollandi for ATM under the Observe & Notify control. Ollandi reads the ATM monitoring platform, switch and transaction state, hardware and software events, site power and carrier status, and security signals through read-only feeds. It does not reset, reboot or command terminals, and it does not dispatch field engineers. The accountable team or vendor carries out every repair.
Cases opened before the threshold
Ollandi tracks availability, dispensing, communications, cash, hardware, software and security as separate dimensions of each terminal's health. When a pattern indicates a developing fault, such as rising dispense rejects or repeated device errors from the same module, Ollandi opens a case marked as a forecast risk before the service threshold is breached. The forecast is labelled as a forecast, with the evidence behind it, so the team can decide whether to act.
Faults explained by their real cause
When a switch link degrades, the affected terminals are correlated into one case explained by the upstream switch condition. When a site runs on generator or UPS, Ollandi records it as normal operation rather than an incident, and raises a case only when backup capacity is at risk. The bank stopped sending engineers to terminals whose problem was upstream.
Missing evidence treated as risk
If a feed goes quiet, the terminals it covers are shown as stale or unknown, never as healthy. A single failed source produces one source-level incident instead of hundreds of false outages.
Security controls watched as service conditions
A failed anti-skimming module or offline camera does not stop a terminal from dispensing, so it rarely reached the top of a queue. Ollandi treats these as exposure: the condition that makes the next loss undetectable. They are now routed to the security team with the same evidence trail as an availability fault.
The operating outcome
The ATM team now hears about developing failures from Ollandi rather than from customers. Field work goes to the terminals that need it, switch and power events are handled by the teams that own them, and gaps in monitoring are visible instead of silently counted as uptime.
Escalation by email was switched on only after the bank validated the workflow and two administrators approved it. Every case still closes on verified recovery.
Operating control: Observe & Notify.
