ATM Management
An AI-based ATM monitoring system that watches your self-service banking fleet around the clock, detecting skimming, forbidden attributes and suspicious behaviour in real time from a central operations view.
Request a demo
Alfabeta's ATM monitoring system is an integrated banking solution that applies computer vision to distributed cash machines and self-service kiosks. It watches each ATM booth for both security and cleanliness, so financial service providers can protect their assets, reassure customers and lift operational efficiency at the same time. Rather than relying on guards at every site or footage nobody reviews until after an incident, the system delivers centralized, integrated and real-time monitoring across an entire fleet.
On the security side, the platform runs banking security analytics tuned to the specific risks ATMs face. It detects RFID-based skimming devices attached at the card slot using machine-learning models, and performs forbidden-attribute detection: if a customer is wearing a helmet, sunglasses, facemask or hoodie, the system alerts them to remove it and can hold the transaction until the face is fully visible. Object and safety-violation detection watches the surrounding area for suspicious behaviour or items, while low-light detection uses computer vision together with infrared (IR) to keep working when lighting is dim or has been deliberately tampered with.
The same infrastructure supports customer verification and onboarding. Alfabeta's e-KYC feature reads the data and photo from a customer's identification card with OCR, then uses face recognition with liveness detection to match the ID photo against a live selfie, requiring a resemblance of at least 75% to count as a match and blocking spoofing attempts. This directly supports Indonesia's OJK e-KYC expectations under POJK 12/2017, under which a financial service provider's own electronic verification method can stand in for in-person facial verification, enabling faster, paperless onboarding.
Operationally, the system also measures people counting and dwelling time, tracks queue length and flags when a customer stays inside a booth too long, and provides a controlled, anti-spoofing attendance record for third-party cleaning and maintenance vendors. Banks and cash-management operators gain a transparent, objective view of vendor performance and booth cleanliness, protect facilities from criminal risk, and oversee hundreds of machines from one dashboard, all of which improve customer security, loyalty and satisfaction.
Tamper & Skimming Detection
Detects RFID-based skimming devices installed at the card slot using machine-learning models, flagging tampering before customers are compromised. Combined with object and safety-violation detection around the booth, it turns the ATM into a monitored asset rather than an unattended risk.
Loitering & Threat Alerts
Observes people and objects in and around the ATM and triggers a local alert whenever suspicious behaviour or items are detected. Low-light detection using computer vision and infrared (IR) keeps this working even when the booth is dark or the lighting has been deliberately tampered with, a common precursor to theft.
Transaction Face Capture
Enforces forbidden-attribute detection so a customer's face is fully visible during use: helmets, sunglasses, facemasks and hoodies trigger an alert and the transaction can be held until they are removed. Captured faces support fraud investigation and dispute resolution.
Centralized Fleet Monitoring
Brings every machine into one dashboard for centralized, integrated and real-time monitoring, so a small operations team can oversee a nationwide fleet. It also measures people counting, dwelling time and expected queue duration, escalating when a customer stays inside a booth too long.
Health & Uptime Monitoring
Tracks camera and site status to catch obscured lenses and offline machines quickly, and identifies necessary activities such as regular maintenance and cleaning. This underpins a transparent, objective rating of third-party vendor performance and booth cleanliness.
Evidence & Audit Trail
Stores video and event logs for compliance, chargebacks and law-enforcement requests, and provides a controlled, anti-spoofing attendance record for cleaning and maintenance vendors. e-KYC verification events, including OCR ID reads and liveness-checked face matches, are retained to support audit and OJK expectations.
Frequently asked questions
How does the ATM monitoring system detect card skimming?
The system uses machine-learning models to detect the presence of RFID-based skimming devices installed at the card slot, alerting the operations team before customers are compromised. This runs alongside object and safety-violation detection around the booth, so both device tampering and suspicious physical activity are caught in real time.
What are forbidden attributes and why are they blocked?
Forbidden attributes are items that hide a customer's face during a transaction, such as helmets, sunglasses, facemasks and hoodies. Because clear face visibility is essential for fraud deterrence and evidence, the system detects these attributes and alerts the customer to remove them, and the transaction can be held until the face is fully visible.
Does the system work in poor lighting or at night?
Yes. Low-light detection combines computer vision with infrared (IR) so the system continues to detect human behaviour and objects in dark conditions. This is important for ATMs in poorly lit locations and for cases where lighting has been deliberately tampered with to enable theft.
How does the e-KYC feature support OJK compliance?
The e-KYC feature reads the customer's ID card data and photo with OCR, then uses face recognition with liveness detection to match the ID photo to a live selfie, requiring at least a 75% resemblance to count as a match. This electronic verification method supports Indonesia's OJK framework under POJK 12/2017, which allows a financial service provider's own electronic verification to replace in-person facial verification, enabling faster, paperless onboarding.
Can one team monitor an entire national ATM fleet?
Yes. The system provides centralized, integrated and real-time monitoring that brings every machine into one dashboard, so a small team can oversee hundreds of locations. Alerts for skimming, loitering, safety violations and excessive dwell time are escalated centrally, and vendor attendance and cleanliness are tracked objectively across the fleet.