Payments

Rails before loans: why Nigeria's push into SME credit depends on payment infrastructure

Embedded finance puts credit inside the payment flows businesses already use. Whether borrowers repay depends on the rails underneath.

Cover: Rails before loans. Why Nigeria’s push into SME credit depends on payment infrastructure. Railway tracks running into the distance.

The CBN launched its Payments System Vision 2028 (PSV 2028) in June. Alongside it, Nigerian fintech is moving from payments into embedded finance: credit offered inside the payment flows businesses already use, priced on their transaction data, and aimed at small businesses that formal lenders have never served.

I think that’s right. Embedded finance is the most practical route to credit for small businesses that we have. I also think we underrate how much it depends on the plumbing underneath it. I’ve spent 15 years building national payments infrastructure, on a platform that moves salaries, pensions and government collections. The lesson I keep coming back to is that a loan book takes on every weakness of the rails it sits on.

Underwriting gets the attention. Collection decides the outcome.

Most of the excitement about SME credit is about underwriting: better data, better models, faster approvals. That matters. But a loan is only as good as your ability to collect it.

Payment data tells you who can repay. Whether they actually do depends on mandates, direct debits and the systems between banks working when they’re supposed to. When a repayment debit fails because something downstream timed out, the borrower hasn’t defaulted. The infrastructure has. It still shows up in arrears, it still triggers a collections call, and it still damages the relationship with a customer who did nothing wrong.

At small volumes that’s noise. Across millions of small loans it skews your risk models and pushes up the price of credit for everyone. So I’d put collection success rate on the same dashboard as approval rate and NPL ratio.

In embedded finance, the rails are the product

Embedded credit works because the payment flow does three jobs at once. It is the distribution: the offer appears on the terminal or app where the business already takes payments. It is the data: the same transactions that show the business is real are the ones that price the loan. And it is often the collection channel: repayment is taken from the money flowing through.

That is what makes embedded finance powerful, and it is also why it is unforgiving. When the rails falter, the offer, the underwriting and the repayment all falter together.

Fast in the middle, fragile at the edges

At the launch, the CBN said most instant transfers across the industry settle in under 10 seconds. The middle of the system is fast. The problems are at the edges: the transfer stuck on “pending”, the reversal that takes days, the dispute nobody owns.

And the Governor has been open that trust in digital payments is still a major challenge. Embedded lending makes this more serious, because now a failed transaction isn’t just an annoyance, it’s a missed repayment.

Three things would help:

  • Measure the whole journey. Each institution reports its own uptime, but a transaction crosses several. A shared, published success rate for the full journey would tell us far more.
  • Make reversals automatic. Failed-but-debited transactions should reverse within a fixed window, the same way across the industry.
  • Use the richer data. The industry is moving to ISO 20022: since 22 November 2025, payment instructions between banks on Swift travel only in that format. The structured data in those messages can make reconciliation, fraud checks and dispute handling much faster, which is exactly what a high-volume credit operation needs.

Lenders see fragments

Most small businesses in Nigeria spread their money across more than one provider. Picture a trader with a POS terminal from one company, two bank accounts, and a wallet she uses to pay suppliers. Each lender sees part of her business and prices the loan on that part. Good borrowers get small limits, and risky ones slip through.

Open banking is the fix. The Governor said at the launch that more than 100 licensed APIs are already available. With the customer’s consent, a lender could look at her whole cash flow and collect from wherever the money actually lands. That matters for embedded finance in particular: without it, a lender built into one platform only ever sees the flows on that platform.

One analyst who went through the 132-page PSV 2028 document for TechCabal counted open finance appearing only once. For the credit ambitions to work, it needs to be a much bigger part of how the vision is implemented.

Compliance has to be automated

As fintechs grow, compliance done by hand doesn’t scale. The direction from the regulator is already clear. According to TechCabal’s reading, the CBN issued baseline standards for automated AML monitoring in March 2026, giving deposit banks 18 months and other institutions 24 months to run machine-learning monitoring. The same analysis says PSV 2028 aims for a machine-readable CBN rulebook and for 90% of institutions to send automated compliance data by 2028.

For anyone building credit products, that means KYC, AML and reporting need to be built into the product from day one, not bolted on after an audit.

What I’d ask for

From builders:

  • Treat reliability engineering as part of your credit function. In embedded finance a failed payment is a failed repayment, so engineering and risk teams should share metrics.
  • Design collections for a world where the borrower’s money sits in several places.
  • Build compliance into the product.

From the industry:

  • Publish end-to-end success and resolution metrics.
  • Standardise reversal timelines.
  • Move faster on consented data sharing, with clear rules on liability and consumer protection.

PSV 2028 has ambitious inclusion goals, and credit is one of the best tools we have to reach them. Embedded finance can put working capital where small businesses already transact, and nothing brings a business into the formal economy faster. But credit built on unreliable, fragmented rails will let down both the businesses that need it and the lenders who extend it.

The order matters. Rails first, then loans.

Sources