Ghana's cross-border payments problem isn't a missing app.

Ghana does not have a shortage of payment apps. It has a shortage of usable dollar liquidity, a correspondent-bank chain that still makes simple transfers feel heavy, and a formal system that informal trade already routes around. Another wallet will not fix that. A rail that shows up inside the bank, can complete in the currency the trade actually needs, and fits the documentation a real shipment requires just might.

I write that from two seats: as someone who has sold payments and digital-asset products to institutions and businesses, including taking a retailer live on crypto payments, and as someone who does commercial work on agribusiness and cross-border commerce in Ghana. The product demos are usually ahead of the operating reality.

The official map is not the route people use

On paper, the region is building the infrastructure this problem is supposed to need. PAPSS, the Pan-African Payment and Settlement System, said in an 11 September 2026 briefing that it now operates in more than 30 African countries, connecting 24 national and regional central banks, more than 200 commercial banks and payment service providers, and 16 switches, with a termination footprint of more than 300 financial institutions. Compared with the same period in 2025, it reported transaction volumes up about 1,000% and values up about 120%. (Afreximbank / PAPSS, 11 September 2026)

Those are the system’s own figures, and large percentage increases can still sit on thin corridors. The commercial fact is simpler. Ghana is on the network. Coverage is not the same thing as a treasury team in Accra using it, every day, for the counterparties they actually pay.

The cost of the older rail is still high. The World Bank’s Remittance Prices Worldwide data for Q3 2025 put the average cost of sending $200 to Sub-Saharan Africa at 8.46%, the highest of any receiving region, against a global average of 6.36%. Banks were the most expensive provider type, at 14.99%. (World Bank, RPW Q3 2025)

When the formal product is slow or expensive, activity does not wait. Chainalysis estimated that Ghana’s crypto economy grew 80% to $5.1 billion in the year to 30 June 2026, and that, across Sub-Saharan Africa, stablecoin remittance value more than doubled to $492 million. (Chainalysis, 29 October 2026) That is not proof that crypto is Ghana’s payments system. It is evidence that people assemble workarounds when the official corridor fails a daily job.

What actually breaks a payment

In cross-border commercial work in Ghana, the failure is rarely “we could not find an app.” It is some combination of these:

• The buyer or supplier needs dollars, and the bank cannot sell them in the size or at the time required.
• The transfer can be initiated, but correspondent steps, compliance queries, or cut-off times stretch a two-day job into a two-week one.
• The counterparty is real, and the paperwork is not in the shape a formal bank process expects.
• The goods are already moving through a channel that is only partly visible to official trade statistics, so the payment path is informal because the trade path is informal.

A new front end does not create FX. It does not complete a bill of lading. It does not make a correspondent bank answer faster. Sellers who begin with the interface are solving the visible part of the problem and missing the constraint.

The legal context should keep payments teams honest. As currently described in the Bank of Ghana’s February 2026 FAQs on the Virtual Asset Service Providers Act (passed by Parliament on 19 December 2025), the cedi remains the sole legal tender, virtual assets are not a substitute for it, and it is not permissible to issue invoices, set prices, or pay wages in virtual assets. (Bank of Ghana) Those are the rules as published at the time of writing. For payments that need to conclude in Ghana, a serious design has to be able to land in cedi, on the books, through authorized providers. That is a Ghana-specific requirement, not a claim that every cross-border payment, in every corridor, must settle in local currency. “Just pay in a stablecoin” is not a Ghana solution, even if some counterparties already transact that way around the formal system.

What I have learned selling the formal rail

When I sold digital-asset and payments products, the deals that worked started with a single live flow, not a platform. A retailer went live on crypto and NFT payments because commercial, payments, and engineering agreed on the first path and someone owned the handoff. Founder-led protocols I helped launch on Base needed the same thing: a route to a working market, not a tour of features.

A competitive win-back into Coinbase Prime taught me a related lesson. The client was leaving over fees. The save was about execution quality and routing, then commercial terms. Cross-border payments in Ghana rhyme with that. Finance leads will say the transfer costs too much. Sometimes they mean the fee. Often they mean the all-in cost: delay, FX availability, uncertainty that the beneficiary is paid, and staff time spent chasing a payment that should have been boring.

Banks and fintechs I sell into now ask a version of the same question, whether the product is embedded finance or a payments rail: can this live inside our existing operations, and what happens when it breaks? If you cannot answer that for a Ghana-to-regional or Ghana-to-US flow, you do not have a product yet. You have a demo.

What actually fixes it

I would not start with a new brand. I would start with four unglamorous pieces of work.

1. Put the rail inside the bank the client already uses. PAPSS only changes a business’s life if a relationship manager, a treasury portal, or a branch process can initiate the payment to the counterparties that matter. Network maps do not move cocoa, cashew, or invoices. Bank implementation does.

2. Treat FX as the product, not a footnote. If the constraint is dollar (or naira, or CFA) liquidity, a faster payment instruction does not help. In many Ghana corridors, the useful product is local-currency initiation, a clear conversion path, and an honest view of when the FX simply is not there. Other corridors may need to complete in dollars or in a regional currency. The point is to design for the currency the trade actually needs, not to assume one settlement currency fits every route.

3. Design for the documentation the trade already requires, not the documentation a pilot prefers. A payments product that only works for clean, fully formal, fully invoiced flows will miss a large share of real cross-border commerce. That is not an argument for weakening controls. It is an argument for helping legitimate firms get the file complete enough to use the formal rail.

4. For Ghana legs, land legally in cedi. Under current Bank of Ghana rules, virtual assets cannot be used to invoice, price, or pay wages, so a serious Ghana design needs a banked, local-currency result through authorized intermediaries when the payment has to conclude there. That does not mean every cross-border payment must settle in local currency. It means the Ghana end of the flow has to fit the rules that apply there. Anything else is a workaround you will eventually have to unwind.

Three things that kill a Ghana payments pilot

1. A corridor no one at the bank can actually operate. The scheme is live in the press release. The core-banking integration, staff training, or beneficiary-bank coverage is not.
2. A first use case that tries to cover the whole map. Accra to one repeat counterparty, in one currency pair, with one cash-out path, is a pilot. “All of West Africa, all clients” is a stall.
3. No owner for exceptions. The first failed or delayed payment will decide whether the treasury team comes back. If compliance, operations, and the relationship manager have no shared playbook, they will return to the informal rail they already trust.

The fix is not more consumer apps, and it is not a speech about financial inclusion. It is making the formal rail cheaper, more available in the currencies people need, and usable by the firms already trying to get paid. Until that happens, the workarounds, including dollar tokens on public networks, will keep doing the job the official system has left open.