Every week, mobile banking customers in Nigeria lose money to SIM swap fraud. The money is gone, the attacker is unreachable, and the operator’s fraud team is left investigating an incident that was preventable — not at the network level, not at the bank level, but at the registration desk.

This post explains exactly how SIM swap fraud works, why it keeps succeeding, and what it actually takes to close the vulnerability.

The Attack, Step by Step

SIM swap fraud is not technically sophisticated. It does not require network intrusion or advanced malware. It exploits a gap in how most operators process SIM swap requests.

Here is how the attack unfolds:

Step 1 — Intelligence gathering. The fraudster collects basic personal details about the target: name, date of birth, phone number. This information is available through social media, phishing, data breaches, and in some cases purchased from insiders. None of it requires hacking a telecom network.

Step 2 — Walk into a retail outlet. The fraudster visits a retail outlet or contacts a dealer and reports the target’s SIM as lost or damaged. On most systems, the agent’s job at this point is to confirm the requester knows the account details. The fraudster does.

Step 3 — Receive the new SIM. On a system that does not require biometric verification for a SIM swap, the agent completes the swap. The fraudster walks out with a SIM card linked to the victim’s MSISDN.

Step 4 — Mobile banking takeover. Within hours — sometimes minutes — the fraudster uses the new SIM to intercept OTPs for the victim’s mobile banking accounts. Password resets. Transaction authorizations. The victim’s phone goes dark. Their account drains.

The entire chain from Step 1 to Step 4 can be completed in a single afternoon.

Why It Keeps Working

The vulnerability is structural, not behavioural. Operators address it through agent training, escalation procedures, and fraud awareness campaigns. None of these close the gap, because the gap is not agent knowledge — it is agent authority.

On most legacy registration systems, the SIM swap workflow requires the agent to verify the subscriber’s identity using information the subscriber provided at registration: name, address, possibly a document number. An agent who has been trained to check these details is still relying on information a fraudster has already obtained.

The harder problem: on systems without biometric verification for SIM swaps, the agent has discretion. And discretion can be pressured, manipulated, or bypassed.

Policy training tells agents what to do. It does not physically prevent them from doing otherwise.

What Closing the Gap Actually Requires

The only control that closes SIM swap fraud at the point of transaction is biometric verification — the same biometric verification that NCC regulations require for new SIM registration.

This is not a new idea. The NCC’s Business Rules for Communications Subscriber Registration (July 2025, 2nd Amendment) are explicit: the same identity verification standard that applies to new registration applies to SIM swaps. Biometric capture. Real-time NIN verification against NIMC. Liveness detection to confirm the person in front of the camera is alive, not a photograph.

Three controls, working together:

Biometric capture with liveness detection. A photograph of the account holder will not pass. The system requires a live facial capture and confirms liveness before the image goes to verification. A fraudster who knows the victim’s name and phone number cannot produce the victim’s face.

Real-time NIMC verification. The captured biometric is matched against the NIMC reference image in real time. The match must score above 70%. Below that threshold, the swap is a hard rejection — no supervisor override, no configuration option, no exception pathway.

Append-only audit trail. Every swap attempt is recorded permanently: the agent, the device, the GPS coordinates, the NIN, the NIMC response, the biometric match score, the AI model version, the routing decision. If a fraudulent swap does occur, investigators have the complete chain of events — not a reconstruction from fragmented logs.

The Question to Ask About Your Current System

When a subscriber walks into one of your retail outlets, reports their SIM as lost, and requests a swap — what does your system require?

If the answer is anything other than live biometric capture and real-time NIMC verification, the SIM swap fraud vulnerability is open. Not potentially open. Open.

The NCC’s July 2025 requirements have raised the floor. Operators running legacy registration systems that handle SIM swaps through document checks and agent discretion are not meeting the current standard — and more importantly, they are not closing the fraud vector that is actively being exploited.

BioSmartX: Biometric-Mandatory SIM Swap

BioSmartX enforces the same nine-step compliance workflow for SIM swaps that it enforces for new registration. There is no shortcut, no fast-lane process, no agent discretion at the biometric step.

A SIM swap request that scores below 70% on the facial match is a permanent hard rejection. Not a flag for supervisor review. Not a case for escalation. A rejection — with no override path, for any user at any level of the system.

Operators running BioSmartX at Glo Nigeria and MTN Côte d’Ivoire have the same biometric gate on SIM swaps that they have on new subscriber registration. The swap process is closed to anyone who cannot produce the account holder’s live biometric.

SIM swap fraud is not going to stop because agents are better trained. It will stop when the registration process makes it structurally impossible to complete a swap without biometric verification of the person standing at the counter.

That is a system decision, not a training decision.

If you want to see how BioSmartX closes the SIM swap vector — specifically, technically, and in production — request a demo.