By Oluwajuwon Omotayo
Nigeria is winning against one kind of fraud and losing badly against another, and the gap between the two is where most people actually lose their money.
In January, NIBSS reported that losses to electronic payment fraud fell to ₦25.85 billion in 2025, down 51 per cent from ₦52.26 billion the year before. Reported incidents dropped to 67,518, continuing a five-year decline from 123,918 in 2021. That is a real achievement, and it came from coordinated industry work: NIBSS says joint measures prevented roughly ₦20 billion in further losses over the same period.
Now set that beside what happened last week. Investors stormed the offices of PXES after the platform stopped processing withdrawals, one of them reporting a loss of ₦209,000. It follows CBEX, which collapsed in April 2025 after promising to double deposits in thirty days, and before that MBA Forex, Racksterli, MMM.
Both of those are fraud. Only one of them shows up in the NIBSS figure.
The difference is who pressed send. The ₦25.85 billion covers money taken out of accounts — compromised credentials, SIM swaps, insider abuse. But when you are persuaded to pay a vendor who never ships, or a P2P counterparty who disappears after you release the crypto, the transaction is not compromised at all. It is authorised, by you, and it settles perfectly.
Nothing in the payment rails treats it as an anomaly, because on every technical measure it is a normal transfer.
NIBSS itself named social engineering as the most prevalent technique in the system. The fraud is migrating towards persuading the account holder rather than defeating the bank, because the bank has become the harder target.
The moment the system leaves you alone
Instant payments in Nigeria reached ₦284.99 trillion in the first quarter of 2025 alone. That speed is the country’s great payments achievement, and it is also the thing that makes this category of loss irreversible. You have a few seconds between reading an account number and losing the ability to do anything about it.
In those seconds, the only question that matters is whether this account has done this before. And that is the one question you cannot get answered.
The intelligence exists. Banks hold it. NIBSS aggregates it across the industry. What it does not do is reach the person about to make the payment. NIBSS has itself warned that declining incident reporting weakens the industry’s ability to track offenders moving between institutions, and that visibility depends on information sharing between banks, telcos, platforms, regulators and law enforcement. That warning is aimed at institutions. The individual making the transfer is not in the list at all.
Why the obvious fix is the wrong one
The instinct is to build a lookup tool that returns a rating. Enter an account number, get green or red. Several products in this market work exactly that way.
That design has a flaw that cannot be engineered around.
A scammer’s first victim looks up the account and finds nothing, because the crime has not happened yet. If the product renders that as green, it has told the person the account is safe at the precise moment it is most dangerous. It did not merely fail to protect them. It replaced their caution with a reassurance it had no basis for. The second victim gets the same green tick, and the third, until someone finally reports.
An empty result carries one piece of information: nobody has reported this person yet. That is not safety, and no interface design makes it so.
The opposite failure lands people in court
So warn more aggressively? Flag anything anyone reports?
Consider what that means in this jurisdiction. In September 2023, Chioma Okoli posted on Facebook that a tomato paste she had bought was too sugary. The manufacturer petitioned the police. She was arrested, arraigned on cybercrime charges, held in a correctional centre, and granted ₦5 million bail. The case has run for over two years. The FCCPC found she had broken no law.
That is what happened to one consumer, over one opinion about tomato paste.
Now imagine a platform publishing thousands of accusations that specific named people took money and did not deliver. The Committee to Protect Journalists reported that at least three Nigerian journalists were detained under the Cybercrime Act in the months after August 2025, despite the 2024 amendment that was supposed to curb exactly this. The ECOWAS Court of Justice ruled Section 24 arbitrary and vague back in 2022 and ordered it repealed. It is still being used.
Any platform publishing fraud allegations at scale is publishing, with everything that word carries. And the harm is not only legal. A single accusation may be true, or it may be a business dispute, a soured relationship, a competitor, or someone who simply copied the wrong account number. Turn one accusation into a public warning and you have handed anyone with a grievance a weapon aimed at a real trader’s livelihood.
The CBEX collapse showed how fast unverified numbers travel even when everyone means well. Early reports put losses at ₦1.3 trillion; later analysis by Techpoint Africa and independent researchers put the real figure closer to $6.1–12 million. The larger number is still repeated today. If a figure that size can circulate uncorrected, so can an accusation against a trader nobody is checking.
What we built instead
That is the box: cannot promise safety, cannot warn on one accusation. It is the constraint we built PeerVet around, and it is worth being concrete about what it forced.
No clean result is ever shown as safety. When there is nothing on an account, PeerVet says nobody has reported this person yet, in those words. There is no green state anywhere in the product.
Evidence has to accumulate before anything becomes a warning. One report flags nobody. Corroboration from unrelated people does. An established reporter carries more weight than an account created an hour ago, multiple reports from one person collapse into a single voice, and accounts that only ever post together count as one. Assembling a crowd to bury someone does not work.
Scraped data can never on its own reach the most serious verdict. Public mentions from forums and blocklists are shown as context. Only accountable in-app reports, tied to a verified phone number, can produce a danger result.
Anyone named can dispute it, and the dispute appears beside the accusation. Filing does not suspend the reports, or a genuine scammer would file once and walk away clean. But the reader sees both. Reporter identities and raw report wording are never published, only neutral summaries and counts.
Good history and bad history are never merged. Fifteen honest trades followed by one exit scam is a known pattern. Those are two separate facts shown separately. A ratio hides the one that matters.
Where coverage is thin, it says so. Our crypto wallet data is strong. Our Nigerian bank account and phone data are much weaker and grow only as people report. The result screen states that rather than implying even confidence.
None of this is as satisfying as a green tick. It asks the person to think rather than obey a colour, and it means we are sometimes slower to warn than we could be. But in a country moving ₦285 trillion a quarter in payments that clear in seconds and never reverse, a product that says “safe” when it means “unknown” is not a neutral failure. It is a participant.
It is live at peervet.ng. It will not give anyone a green tick, and that is deliberate.
Oluwajuwon Omotayo is CTO at Sage Grey Technologies and founder of Ginux Technologies.