The Core Debate
When financial institutions build or upgrade their transaction monitoring infrastructure, one of the most consequential decisions they face is whether to screen transactions in real time or to process them in batches at intervals. Both approaches have legitimate use cases, and the right answer depends on the institution's specific risk profile, transaction volumes, and operational constraints. But in the context of African fintech, where fraud moves fast and regulatory scrutiny is increasing, the case for real-time processing has become substantially stronger.
Batch processing, where transactions are collected and screened against watchlists and rules at regular intervals, was the dominant model for a generation of banking compliance systems. It made sense when transaction volumes were lower, computing resources were expensive, and the regulatory expectation was daily reporting rather than real-time detection. That environment no longer exists in most of Africa's major financial markets.
Why Batch Processing Falls Short in High-Velocity Markets
The fundamental problem with batch processing in a high-velocity fraud environment is timing. If a fraud event occurs at 9am and the batch runs at midnight, there is a fifteen-hour window in which the institution is blind to what is happening. In Nigeria's mobile money ecosystem, where mule networks can move funds through multiple institutions in minutes, a fifteen-hour detection window is not a compliance gap. It is an open door.
Batch processing also creates a particular problem for account takeover fraud. When credentials are compromised and funds are moved, the window between the fraud event and the victim's report is often a matter of hours. An institution running nightly batch processes will typically discover the fraud at the same time as the customer, providing no opportunity for pre-emptive intervention.
The Case for Real-Time Screening
Real-time screening evaluates transactions at the moment they occur. This enables institutions to block suspicious transactions before they complete, trigger step-up authentication for high-risk activities, and flag patterns for immediate investigation rather than next-day review. As we have explored in our analysis of continuous risk monitoring, the shift from periodic to continuous evaluation fundamentally changes what an institution can detect and when.
The objection to real-time processing is usually latency. Adding a screening step to every transaction adds time to each payment. In practice, well-designed real-time systems add milliseconds rather than seconds, and the latency is imperceptible to the customer. The engineering challenge is real but solvable.
A Hybrid Approach for Resource-Constrained Institutions
Not every institution can move to fully real-time screening immediately. For institutions with legacy infrastructure or constrained technology budgets, a hybrid approach makes sense. High-value and high-risk transactions are screened in real time. Lower-risk, lower-value transactions are screened in near-real-time batches running every few minutes rather than once daily. This approach captures most of the fraud detection benefit of real-time screening at a fraction of the infrastructure cost.
Remllo's WatchTower platform supports both real-time and configurable near-real-time screening, allowing institutions to match their monitoring intensity to transaction risk without requiring a full infrastructure overhaul.
Regulatory Expectations Are Moving Toward Real-Time
The CBN's evolving guidance on transaction monitoring reflects a regulatory direction of travel toward real-time or near-real-time detection. Institutions that can demonstrate real-time monitoring capabilities are increasingly differentiated in regulatory examinations from those relying on legacy batch systems. The shift is gradual but the direction is clear.
