What, Exactly, Will Make a Bank a Bank in 2035?

Ethel D. Cofie

African banks are making expensive choices about their future. They are modernising technology, investing in AI and payments, partnering with fintechs and telcos, expanding digital distribution and buying new capabilities. But underneath those decisions sits a more fundamental question: which parts of a bank will actually matter most ten years from now?

By 2035, a customer might save through one interface, borrow through another, pay through a third and invest through a fourth without thinking about which regulated institution sits underneath each service. A bank may provide the balance sheet without owning the interface. It may provide the product without owning the distribution. It may hold the deposits while another company has the customer’s attention, transaction data and first opportunity to sell the next financial product.

What can a bank stop owning without becoming less valuable?

What must it continue to control?

And how many components of banking can become interchangeable before the bank itself becomes interchangeable?

My argument is that the strategic challenge facing banks is shifting from owning the banking value chain to deciding which positions within it they cannot afford not to control.

This does not necessarily mean banks should own more. A bank may become more valuable by owning less where technology, infrastructure or distribution can be sourced more efficiently elsewhere. But there is a line between giving up activities that no longer differentiate the institution and giving up the customer relationship, information, economics or risk capability on which its future value depends.

The strategic problem is knowing where that line is.

A Bank Was a Bundle

We talk about banks as if they are one thing. They never were. A bank has traditionally bundled together deposits, lending, risk, payments, financial products, distribution, customer relationships, information and the technology required to make the whole thing work. Regulation, capital and institutional trust sat around that bundle.

For a long time, these functions largely travelled together. Distribution was physical. Information was expensive. Technology was costly and often proprietary. Regulation placed responsibility on an identifiable institution.

That has changed. A telecommunications company can provide the interface while a bank provides the loan. A fintech can originate the customer while another institution provides the capital. A merchant platform can hold valuable information about the customer while technology and infrastructure come from still other providers.

The familiar discussion about the “unbundling” of banking explains how this becomes possible. The more interesting question now is what happens to the strategic position of the bank once it becomes normal.

Whose Customer Is It?

Kenya offers perhaps the clearest illustration. Safaricom has more than 40 million active M-PESA customers. Financial products inside that ecosystem are provided with banks including NCBA and KCB. A customer can borrow, transact and repay without entering a conventional bank-controlled interface. In the year to March 2026, Fuliza alone disbursed KES 1.47 trillion to 17.7 million distinct users.

Safaricom has the attention and everyday interface. Banks provide funding, regulated financial capabilities and risk-bearing. Both participate in the economics.

So whose customer is it?

Perhaps “customer ownership” is no longer a useful enough concept.

Who has the customer’s attention?

Who holds the deposits?

Who sees the transactions?

Who makes the credit decision?

Who provides the capital?

Who has permission to offer the next product?

Those positions used to travel together. Increasingly, they do not.

Ghana gives us another version of the same problem. Mobile money transaction values reached GH¢4.54 trillion in 2025, according to Bank of Ghana data. A customer may experience the relationship almost entirely through a mobile-money provider while the float supporting those wallets ultimately sits in trust accounts within the banking system. Interoperability has meanwhile made it progressively easier for value to move between wallets, banks and other institutions.

The mobile-money provider may have frequency and customer attention. Banks hold the underlying float. Shared payment infrastructure connects the institutions. The regulator determines the architecture within which they operate.

As moving money itself becomes easier, the strategic value may increasingly sit around the transaction: who understands the customer, who sees the merchant activity, who can extend credit and who gets the next opportunity to offer a financial product.

This is not a story of telcos winning and banks losing. The financial relationship itself is being distributed across institutions.

Follow the Value

This distinction matters because the institution with the most customer interactions does not necessarily have the most attractive economics.

African banks remain highly profitable. Industry revenues were estimated at approximately US$107 billion in 2025 and average return on equity at around 17%, substantially above the global banking average. Deposits remain valuable sources of funding and lending remains a major revenue pool.

So losing the interface is not necessarily the same as losing the economics. But neither should we assume that the economics will remain where they are today.

Payments are a useful example. A payment generates a transaction, but the economics surrounding it may be more important. Frequent payments generate information. Merchant payments create a view into the merchant’s business. That information can improve a credit decision. Credit can deepen the relationship, which can attract deposits, insurance, foreign exchange or other services.

Payment may be the doorway rather than the room.

And this leads to what I think is one of the more important questions for banks.

When a capability becomes interchangeable, what becomes more valuable around it?

Core banking technology can be bought from multiple providers. Cloud infrastructure is available to competitors. Sophisticated AI capabilities will increasingly be accessible across the industry. Payment infrastructure is becoming more connected and shared.

None of this makes those capabilities unimportant. But it can change where differentiation sits.

If every bank can access increasingly sophisticated AI, proprietary information, judgement and execution may become more valuable. If payment infrastructure becomes common, transaction information and merchant relationships may become more valuable. If products can be manufactured by one institution and distributed by another, permission to place the next product in front of the customer may become more valuable.

Interchangeable does not mean unimportant. It means we need to look for where the scarcity has moved.

Distribution Is No Longer One Thing

For much of banking history, distribution was easy to recognise. It was the branch network.

Today it might be a bank app, an agent, a merchant, a mobile-money account, accounting software used by an SME or a fintech placing a bank’s loan inside its own product.

Equity Group and NCBA illustrate two different possibilities. Equity has moved the overwhelming majority of its transactions outside branches while largely keeping those interactions within its own ecosystem. NCBA has used telecommunications distribution through products such as M-Shwari and Fuliza to reach customers at enormous scale.

One model retains the interface. The other is willing to use somebody else’s.

Both can work.

So the useful question is not simply whether a bank owns distribution. It is what the bank receives from that distribution.

Does it receive deposits? Transaction information? Credit origination? Merchant relationships? Low-cost customer acquisition? Permission to sell the next product?

Or does it receive mainly volume while another institution accumulates the more valuable parts of the relationship?

Watch What Banks Do

Capital allocation may tell us more about the future of banking than predictions do.

Standard Bank processed more than R164 trillion across approximately 2.3 billion payments in 2025. Payments at that scale are not simply transactions. They sit alongside deposits, working capital, foreign exchange, collections, trade and other financial relationships.

Ecobank presents a different strategic asset. Its regulated presence across more than 30 African markets gives it something that cannot be reproduced simply by purchasing better technology: the ability to connect customers, businesses and financial institutions across multiple jurisdictions.

That becomes particularly interesting if technology itself becomes easier to acquire. What happens to the relative value of regulated reach when sophisticated technology becomes increasingly common?

Elsewhere, African banks are buying fintech and payments businesses, building digital ecosystems and partnering with telecommunications companies and technology providers.

The individual strategies differ, but the underlying decisions are similar:

What must we control?

What can we safely allow somebody else to provide?

Look at what banks build, what they buy, what they partner for and, perhaps most importantly, what they refuse to relinquish. Those choices tell us where they believe future value will sit.

The Decision Looks Different From the Boardroom

This becomes much less theoretical at board level.

A partnership that produces rapid customer growth may look attractive until the board asks which institution retains the transaction information, who has permission to offer the next product and whether the bank is strengthening its position or simply providing regulated balance-sheet capacity behind somebody else’s relationship.

An embedded lending partnership may generate impressive volumes while leaving the bank carrying credit risk on customers it did not acquire and relationships it does not control.

The issue is therefore not simply whether an investment or partnership meets its business case. The board needs to understand what position in the financial relationship the decision is buying, protecting or potentially surrendering.

If another institution owns the interface, what does the bank receive in return?

If the bank carries the credit risk, what information and control does it retain?

If customer activity moves through a partner, who accumulates the data and the right to make the next offer?

How would we know if the bank were gradually becoming a balance-sheet provider for somebody else’s customer relationship?

And if that were deliberate, are the economics good enough to make that an attractive business?

That last distinction matters. Becoming a balance-sheet provider is not necessarily strategic failure. At sufficient scale, with attractive risk-adjusted returns and funding economics, it may be an excellent business.

The board question is whether that position is deliberate and economically attractive, or the unintended result of a series of individually sensible decisions.

What Proves Harder to Give Away?

Not everything is becoming interchangeable.

Some of the least visible parts of banking may prove the hardest to reproduce.

Somebody ultimately has to fund the loan and absorb the loss if it fails. Deposits remain an important source of funding. Regulated financial intermediation is capital intensive. Complex corporate relationships involving trade finance, treasury, foreign exchange, guarantees and liquidity management are considerably harder to reproduce than a digital interface.

Risk judgement may also become more important as the technology supporting it becomes more common. If sophisticated models are available to everybody, the quality of the information feeding those models, institutional experience and the willingness to put capital behind decisions may become more important rather than less.

Then there is trust. I would be careful about treating trust as something banks automatically own. African consumers have demonstrated enormous trust in mobile-money providers, while banks themselves have sometimes lost public confidence.

Trust may belong less to an institutional category than to whichever institution repeatedly does what it promised to do, particularly when something goes wrong.

This creates an interesting tension. Some of the most visible components of banking are becoming easier to separate from the bank, while some of the least visible may prove the hardest to move.

What Can the Bank Afford Not to Own?

Perhaps the successful bank of 2035 will own considerably less than the successful bank of 2025.

It may not manufacture every product, originate every customer, own every distribution channel, build every piece of technology or control every customer interface. That does not necessarily make it weaker. Owning capabilities that no longer differentiate the institution can consume capital, technology investment and management attention without creating equivalent strategic value.

But there is a dangerous opposite.

A bank can outsource technology, partner for distribution, use common infrastructure and place products inside somebody else’s customer experience through a series of individually sensible decisions. At some point, it may discover that it has also surrendered the information, relationship, pricing power or strategic position that made its economics attractive.

That is the line bank leadership needs to find.

There will not be one answer for every institution. A mass-market retail bank may need to control something different from a corporate bank. A bank operating across multiple African markets may derive advantage from something a domestic institution cannot reproduce. An SME-focused bank may place particular value on transaction information and merchant relationships.

Which is why I am sceptical of any single model of “the bank of the future”.

The more useful question may be: what must remain distinctly ours?

Back to 2035

By 2035, the customer relationship may be distributed across several companies. Products may move easily between institutions. Technology may be widely available. Distribution may belong to somebody else. Banks, fintechs and telecommunications companies may offer increasingly similar financial products.

And yet somewhere inside that system there will still be institutions holding deposits, allocating capital, carrying risk and accepting regulatory accountability for financial promises that may stretch decades into the future.

The strategic challenge is already here because African banks are making decisions today about technology, distribution, partnerships, payments and acquisitions that will determine which positions they occupy in that future system.

So I return to the questions at the beginning: