Catching Fraud in the Act
A Nigerian AI startup is taking on the $5 trillion global fraud problem
Hi there,
Welcome to the 40th edition of Next Capital, where we help you find Africa’s most promising startups, before they get big.
In our last edition, we wrote about the startup breaking the $4.6 trillion spreadsheet economy. If you missed it, you can catch up here 👇🏾
90% discount on partnerships till March
Get in front of 11,000 investors, founders, execs, and tech enthusiasts in Africa. Find out more here.
Startup: Scrub.io
Location: Nigeria
Ask: Raising a $500k pre-seed round
The world’s GDP is $117 trillion, and every year, trillions of dollars are lost to fraud. In 2024 alone, $5 trillion was lost to fraud.
This happens across different levels - payments, identity theft, insurance scams, and investment scams. It’s an economy in itself, one that’s as old as society.
Since its introduction in 2010, digital payments in Nigeria have exploded, reaching quadrillions of naira. In that time, fraud has also grown.
In 2024, Nigerian financial institutions lost ₦52.26 billion ($132 million) to fraud, nearly triple what they lost in 2023. This doesn’t mean banks aren’t fighting fraud. In fact, fraud cases dropped 31% between 2023 and 2024. Each new case just causes more damage.
This means fraudsters are getting fewer, but smarter. And that’s because they take advantage of a system banks across the world use to fight fraud: the rule-based system.
It goes something like this:
If the transaction amount is < ₦500,000 AND it’s a new device, flag it
If the location suddenly changes from Lagos to Vietnam within 30 minutes, flag it
If the card is used more than 10 times in 1 minute, block it temporarily
This system is fast, explainable, and regulator-friendly, but if fraudsters are great at anything, it’s adaptation. Cybercriminals sometimes bend these rules to make fraud possible. Like making a periodic stream of $250 transfers instead of a single $5,000 one.
Last week, Next Capital caught up with Precious Ajayi, an ex-Microsoft tech lead from Nigeria, who saw this firsthand. In 2024, his friend clicked a phishing link that gave hackers access to his account, allowing them move $4,000 out of it at 3 am.
Despite reaching out to his bank for a resolution, he never got his money back. This showed Precious that rule-based systems missed the biggest fraud tell: human behavioral patterns.
So, he teamed up with a friend who had led the product team for the fintech arm of Access Bank, Nigeria’s biggest bank by assets. They both knew the problem well, and together they started building Scrub…
An AI agent that screens transactions for fraud
Scrub is a machine learning platform that prevents fraud before it happens.
Unlike rule-based systems, Scrub uses machine learning models to learn how each user actually spends money: location patterns, spending habits, transaction timing.
Then, when a pattern that’s way outside a user’s behavior is about to happen, Scrub flags and stops it.
For instance, if someone who spends 150,000 naira monthly in Lagos suddenly tries to move 900,000 naira in Kano at 3 am, Scrub stops it - before it clears.
Scrub offers its product as an API that operates across three functions:
Political Exposed Persons (PEP) screening for customer onboarding
Real-time fraud monitoring that flags suspicious transactions
Regulatory reporting that gives fintechs insights into fraud patterns
The first helps banks monitor user risk levels from a political standpoint, the second stops fraud in its tracks, and the third helps build a blacklist that fintechs and banks can use to identify bad actors.
Scrub is currently securing the necessary data privacy licenses from NDPC (Nigerian Data Protection Commission) and the ISO (International Standards Organization).
It’s building its platform to be country-specific, so as it expands, it’ll train its platform based on local transaction data.
Money Talk
Scrub charges $0.01 per API transaction, with volume-based pricing for high-transaction clients.
They’re currently in beta with five pilot users: a Nigerian commercial bank, a halal neobank, and a bookkeeping platform, among others.
Some banks have already expressed interest in a white-label solution, opening up possibilities for licensing revenue.
Scrub has raised $25,000 to date and is in conversations to raise a $500,000 pre-seed round.
The Big Picture
Fraud is no longer just a banking issue. As commerce, work, and identity systems move entirely online, it’s now an infrastructure hurdle for the internet.
Flutterwave, Nigeria’s biggest fintech powering thousands of merchants, lost ₦11 billion in a security breach in April 2024.
Scrub is serving as critical infrastructure here by enabling anonymous data sharing across its network. So, if someone commits fraud at OPay, the pattern gets flagged system-wide, making it harder for bad actors to hop between platforms.
There are local competitors like Archer, YouVerify, and Prembly, but they either focus on one attack surface, like identity, or are building a solution that isn’t AI-powered.
Globally, though, competition is fierce. Global spending on fraud solutions is projected to grow from $34 billion in 2025 to $123 billion by 2033.
Companies like Feedzai, Riskified, and Kount offer ML-based fraud detection and behavioral monitoring, right up Scrub’s alley.
Another hurdle is the procurement cycles at banks, which are notorious for being conservative. Institutions take 6 - 18 months to decide, and even then, tend to trust companies with long track records.
If Scrub cracks the hurdle of getting into enough institutions, the network effects start to kick in for its users, making them less likely to switch.
Do you think Scrub is well-placed to win this market?
Want to be featured in Next Capital? Drop us a line.
Until next week! 🫡
We’re always coordinating participation across active opportunities within the Next Capital pipeline, particularly for investors who can back startups.
If you’d like to be considered for current or upcoming opportunities, let us know, and we’ll follow up where there’s alignment.
Enjoyed reading this? Forward it to someone you think would enjoy it as well.