By Rizmo

Estimated reading time: 10 minutes.

Your money has left your account. The recipient has received nothing. The bank tells you to call the wallet provider. The wallet provider tells you to call the bank.

Who takes responsibility?

That illustrative payment failure puts a difficult question at the heart of Africa’s digital finance agenda: what does progress mean when the customer cannot get a clear answer?

The growth is real. In Sub-Saharan Africa, 58 percent of adults had a financial account in 2024, up from 49 percent in 2021, according to World Bank data published in 2025. The measure covers people aged 15 and older and includes accounts at financial institutions or mobile money providers. World Bank financial inclusion findings.

Mobile money is central to that story: 40 percent of adults aged 15 and older in the region held a mobile money account in 2024, compared with 27 percent in 2021. These account categories overlap; the figures should not be added together. World Bank analysis of digital financial inclusion.

The next investment decision should ask how reliably those accounts serve people. For central banks, governments, banks and financial technology companies, trust needs to become a measurable delivery priority.

For a small business, a delayed payment can interrupt a delivery. For a household, an unexplained charge can reduce the money available for food or transport. For a new customer, a failed first transaction can undermine confidence before a service becomes useful. These are reasons to assess the entire customer journey, from registration to payment and complaint resolution.

The practical agenda starts with six questions. Each asks institutions to make a decision, assign responsibility and demonstrate that the service works.

Can a payment cross networks without losing accountability?

A connection between two systems does not settle who handles a missing payment.

The Bank for International Settlements identifies three dimensions of interoperability: systems must exchange information, interpret it consistently and work under compatible business arrangements. Interoperability between payment systems across borders.

Operators should agree on access, pricing, settlement, transaction status and the handling of failed transfers. Customers should see the full charge before paying and have one clear entry point for help afterwards.

Before expanding a service, follow an ordinary payment from sender to recipient. Then test what happens when it fails. Who investigates? Who updates the customer? Who authorises a reversal? How long does each step take?

If those answers are unclear, the service needs more work.

Consider a hypothetical shopkeeper receiving a payment from a customer on another network. The customer’s screen says the transfer succeeded, while the shopkeeper sees no credit. Should the goods be released? Who can confirm the transaction’s true status? A well-designed service should give both parties consistent information and a dependable investigation route.

Payment operators should test such cases with participating institutions before launch. Their shared rules should cover duplicate messages, delayed confirmations and reconciliation when records disagree. Commercial arrangements also matter: access charges and participation requirements should be transparent enough for a new provider to assess whether joining is viable.

For regulators, the question is whether the scheme’s rules protect reliability and fair participation. For providers, it is whether staff can resolve a problem without passing the customer between organisations. Track failure rates alongside the time customers spend waiting for an answer.

What happens after the innovation pilot?

A promising product needs a credible route to market. Regulators need enough evidence to decide whether it can operate safely.

A regulatory sandbox, where firms test services under controlled conditions, can help resolve uncertainty. The World Bank cautions that counting admitted firms alone says little about a sandbox’s achievements. World Bank lessons from regulatory sandboxes.

Every test should therefore start with a specific question, clear customer safeguards and an exit decision. What evidence would justify authorisation? What would require changes? What would stop the service?

Regulators should publish the pathway and resource the supervision behind it. Requirements should reflect the activity’s risks and potential harm. Where ordinary licensing or clearer guidance can resolve the issue, use that route.

A published framework is a starting point. Its value depends on the decisions and customer outcomes it produces.

Suppose a company proposes a new way for small traders to receive payments. In a hypothetical pilot, the regulator could limit the number of customers and transaction values while testing a specific concern, such as whether funds remain protected if the provider stops operating. Customers would need to understand the test’s limits, how to withdraw their money and where to complain.

Before the pilot begins, agree on the evidence required for the next decision. A successful technical demonstration would not, on its own, answer questions about safeguarding money, handling disputes or supervising the business at scale.

The process should also leave room for a negative finding. A test that reveals an unsafe design can be valuable if it leads to correction or an orderly stop. Publish appropriate lessons without exposing customer information or confidential commercial details. This helps future applicants understand what acceptable delivery requires.

Who gets excluded before the first transaction?

Consider the smartphone assumption. In Sub-Saharan Africa, just 32.9 percent of adults aged 15 and older owned a personal smartphone in 2024, according to the World Bank. A smartphone-only service therefore sets a substantial access condition before anyone can use it. The Little Data Book on Financial Inclusion 2025.

Identity checks can create another barrier. Electronic customer identity verification needs reliable records, lawful access and a practical route for correcting errors. The World Bank’s identification guidance warns that authentication methods can disadvantage people with limited literacy or connectivity. Credentials and authentication guidance.

Governments and providers should build assisted registration, suitable basic-phone channels and alternatives when authentication fails. Share only the personal information needed for the task.

Measure rejected applications and investigate the reasons. How many eligible customers abandon registration? How many cannot correct an identity error? Those are service failures worth fixing.

A hypothetical rural customer might travel to an agent only to discover that the name on an identity record differs from the name held by the provider. Repeatedly asking the customer to return does not resolve the mismatch. A useful process identifies the responsible institution, explains the correction steps and gives the customer a way to check progress.

Providers should distinguish an identity problem from a connectivity problem, an unsuitable document requirement or a poorly designed registration screen. Each calls for a different response. Where local law permits proportionate checks for lower-risk services, explain the limits clearly and provide a route to additional services when the customer meets further requirements.

Privacy deserves the same practical attention. Decide who can access verification records, what information is retained and how mistakes or misuse can be challenged. A successful registration should give customers confidence about both access and the treatment of their information.

Would your service survive a partner’s failure?

A bank or payment company may rely on a telecommunications network, cloud service, identity provider and several technology suppliers. A disruption at one partner can interrupt the customer’s entire journey.

The Bank for International Settlements’ Financial Stability Institute highlights the risks created when institutions depend on a small number of third-party providers. Its banking guidance offers a useful reference, while obligations for other providers depend on local regulation. Sound management of third-party risk.

Map critical dependencies. Test recovery with the partners involved. Agree on incident reporting, customer updates and workable exit arrangements.

Then ask a direct question: when did we last demonstrate that we could restore this service?

A realistic exercise should include an unavailable critical supplier, incomplete transaction records and customers seeking urgent updates. Test who has authority to pause a service, activate recovery arrangements and communicate with supervisors. Confirm that emergency contact details work outside normal office hours.

Recovery also requires checking the money. Restoring an application is insufficient if balances, pending transfers and transaction histories remain inconsistent. Institutions should reconcile records and explain what customers can safely do while the investigation continues.

For a smaller provider, proportionate preparation may begin with a clear dependency map, tested backups and agreed incident procedures. For a major operator, the exercise may require several institutions and supervisors. The scale can differ; responsibility should remain clear. Senior leaders should review unresolved weaknesses, assign corrective work and check the evidence that it has been completed.

Can customers protect themselves and get help?

Clear charges, fraud awareness and usable complaints channels should be part of the product from the beginning. Teach customers how to verify recipients, protect personal identification numbers and report suspicious requests. Test whether they understand.

The World Bank recommends accessible complaint channels, communicated resolution timelines and external redress options. Providers should make those routes visible and track unresolved cases. Complaints handling and dispute resolution.

Women also need a deliberate place in service design. The Mobile Gender Gap Report 2026 estimates that, in Sub-Saharan Africa in 2025, women aged 18 and older were 26 percent less likely than men to use mobile internet. This is a relative gap, not a difference of 26 percentage points. Mobile Gender Gap Report 2026 regional findings.

Ask women, rural customers and people with disabilities to test the service. What is unaffordable, confusing, inaccessible or unsafe? Use their answers to change delivery, then check whether usage improves.

A hypothetical customer using a shared phone may need discreet notifications and a safe way to recover access. Someone with limited literacy may need an agent to explain charges in a familiar language. A customer with a visual impairment may need a registration process that works with assistive technology. These needs should be explored with the people affected rather than guessed in a design meeting.

Consumer education should complement provider safeguards. Customers should not have to understand the entire payment infrastructure to report a missing transfer. Train agents and support staff to recognise problems, record complaints and explain the next step without making promises they cannot keep.

Complaint data can also reveal where design is failing. Review recurring unexpected charges, confusing messages and repeated transfer disputes. Check whether a falling complaint count reflects fewer problems or greater difficulty reaching support. The quality of the answer matters as much as the number of cases closed.

Who owns the next improvement?

Start with a focused 90-day review of a few important customer journeys. Establish the baseline. Identify the failures that cause the greatest harm. Give each improvement a named owner, funding and a deadline.

Track successful payments, recovery times, unresolved complaints, rejected registrations and sustained use. Set targets from local evidence; Africa’s markets have different constraints and capacities.

The test is straightforward: can more people use the service safely, understand what it costs and get help when it fails?

That is a result worth building for.

A workable first month could focus on selecting the journeys, agreeing responsibilities and establishing reliable baseline measures. The second could test the highest-priority changes with customers and participating institutions. The third could assess results, document remaining risks and decide what should expand, change or stop. This is a proposed sequence, not a universal timetable: legal changes or major infrastructure work may require longer.

Keep the review disciplined. A steering meeting should resolve decisions that individual institutions cannot settle alone, rather than collect presentations about activity. Bring evidence of what improved, what remains unsafe and who needs to act next.

For leaders approving the next investment, the final question is simple: what will a customer be able to rely on that they cannot rely on today?

Rizmo is Primproa’s research and writing contributor.