A business owner with more orders than the company can fulfill may see little reason to build a website. A financial institution overseas may still need information it can’t find. The same business can be familiar to its customers and difficult for a banking partner to assess.
Hammed Afenifere, co-founder and CEO of Oneremit, a cross-border payments startup, encountered that tension while speaking with a Nigerian business owner. The owner described about 30 years of trading, yet saw little commercial value in creating an online presence.
That encounter informs Afenifere’s work on AI-assisted compliance and credit assessment. His concern is that the information a financial system expects may differ from the records a business has reason to create, leaving meaningful activity out of view.
Credibility with customers is not the same as evidence for a bank
Afenifere initially encouraged the owner to create a website for exposure and credibility. In his account, the owner saw existing sales as evidence of customer confidence, while orders already exceeded what the business could fulfill. More exposure didn’t address an immediate commercial problem.
The disagreement reveals two different uses of information. For the owner, credibility meant customers continuing to buy. For an overseas financial partner, those relationships weren’t necessarily visible or assessable. A website might help describe a business to someone outside its trading network, even when attracting additional customers wasn’t the owner’s priority.
The payment need made that distinction consequential. Afenifere describes the owner as seeking to stop paying suppliers in China through a middleman. Oneremit’s team submitted the owner’s documents to its banking partner for the proposed payment. The business wasn’t looking for a new reason to transact. It wanted a different way to get money to an existing supplier.
“The banking partner couldn’t assess the individual because there wasn’t enough of an online footprint to complete their due diligence,” Afenifere says. Banks often rely on publicly available information as part of their compliance checks to understand who they’re dealing with and determine whether they can process a payment for a customer”
The immediate question was payment eligibility: whether the banking partner could settle the transaction for the business. The documents Oneremit submitted hadn’t resolved the partner’s need for information. The owner’s confidence and the partner’s uncertainty could coexist because they were approaching the same transaction with different evidence available to them. Customer trust served the owner’s day-to-day trade; making that business assessable to an overseas financial institution presented a different problem.
What local context changes in AI-assisted compliance
Within Oneremit, Afenifere also works on the company’s own use of AI for compliance, helping assess whether activity meets financial rules and obligations. He says part of his role is making that technology useful for African businesses transacting with Western counterparties. The challenge he identifies is insufficient information about the business being assessed, particularly when its operating practices differ from those a system is equipped to understand.
“Internally, we use AI for compliance,” Afenifere says. “One of the setbacks of using AI for compliance is that these AI systems are trained on Western datasets, so to speak. They’re trained from the perspective of someone in the US or Canada, basically, the Western world.”
His concern is that assumptions drawn from one setting can shape what a system expects to find in another. If a business’s activity is poorly represented in the information available, the assessment has less to work with. Understanding why that information is limited matters to interpreting the absence: it may reflect how the business operates, not how much business it does.
Afenifere says his team takes those operating circumstances into account in its internal assessment work. For Oneremit, that means considering how its customers operate when interpreting the information available for a payment assessment. His emphasis is on considering what the available records capture and what they leave out, instead of treating a familiar pattern of documentation as the only way a business can operate.
He also says he works with his team on obtaining licenses where possible. His local-context work sits within that commitment to compliance. The practical challenge is to assess a business for the payment it needs without treating the records it has little commercial reason to create as a complete picture of its activity.
A debit history answers a different question from a credit history
Afenifere carries that thinking into Oneremit’s credit financing, where assessing a business means considering its ability to repay a loan. He says he designed the company’s credit-assessment approach around businesses’ own operating circumstances.
To explain that approach, he points to his own preference for using debit. He describes his credit score as average at best and imagines how a bank might limit what it would approve for him, despite his purchasing activity.
“That means if I approach a bank for payment, they’re probably not going to approve me for a bigger limit,” Afenifere says. “But if you check my debit history, you’re going to see that I’ve made a lot of purchases using my debit card because that is the mentality.”
His example draws attention to activity that a borrowing history alone wouldn’t show. A customer who prefers debit may make substantial purchases without building a comparable record of borrowing. The two histories describe different parts of that person’s financial life.
In designing Oneremit’s approach to credit financing, Afenifere says he starts with the circumstances of the businesses it serves. His debit example explains why: a familiar financial record may capture only part of how someone uses money. He wants the same attention to a business’s actual activity to inform how Oneremit assesses it for financing.
Assessing the business behind the records
For the Nigerian business owner, the immediate need was a way to pay an overseas supplier without a middleman. Oneremit’s involvement brought that practical need up against the banking partner’s difficulty assessing a business with little online history.
That tension helps explain Hammed Afenifere’s focus at Oneremit: bringing the circumstances of the businesses it serves into its work on AI-assisted compliance and credit assessment.
This story was distributed as a release by Jon Stojan under HackerNoon’s Business Blogging Program.