Why Africa’s cash mirage needs vertical software rather than more roadside agents.

Technology conferences in London and San Francisco routinely celebrate Africa as the pioneer of the cashless economy, a continent that bypassed brick-and-mortar bank branches to put a digital wallet in every pocket. At the top of that narrative sits the newly consolidated Vodacom Fintech Group. By March 2026, the group united M-Pesa across East Africa, Vodacash in Egypt, and its South African financial services business into a single operation serving more than 103 million customers and moving $600 billion in annual transactions. Walk a few streets away from the glass towers of Nairobi’s Upper Hill, Lagos’s Victoria Island, or Dar es Salaam’s central business district, however, and between 60 and 90 per cent of daily transactions still happen in paper notes.

Two decades of building vast agent networks did not digitize commerce. Consumers understand mobile money well enough, yet the industry built an architectural detour. Because mobile money operators and agent-banking startups chased cash-in and cash-out (CICO) commissions, they ended up constructing a network of human ATMs that sped up the circulation of physical notes. With CICO withdrawal fees now shrinking across mature markets, commercial banks, telecoms operators, and fintech chief executives face a narrow window before 2030 to replace over-the-counter withdrawal tolls with B2B operating software and industry-specific payment workflows.

Years ago, Tayo Oviosu, founder and chief executive of Nigerian payments company Paga, sat across from officials at the Bill & Melinda Gates Foundation and made an unpopular prediction. Development financiers at the table expected physical mobile money kiosks to serve as a brief bridge lasting three or four years before Nigerians moved to pure digital transfers. Oviosu argued the agent network would run for twenty years. Seventeen years into Paga’s operations, hundreds of millions of venture and corporate dollars have blanketed Nigeria with point-of-sale (POS) terminals and roadside umbrellas, yet roughly 80 per cent of transactions in Africa’s largest economy still settle in physical cash. Across every competing network in Nigeria today, the single most common transaction at an agent kiosk is a customer walking up to a POS terminal to withdraw paper naira so they can spend paper naira down the street.

“For all the money spent on the agent business and building out that agent network across all competitors including us, the biggest transaction at the agent is people going to a point-of-sale machine to get cash. They want to get cash to spend cash. It has actually driven the use of cash; it has not digitised the economy.”

Tayo Oviosu, Founder and CEO of Paga

That same pattern holds across the continent. GSMA industry benchmarks cited by telecoms executives show that outside Kenya, ecosystem transactions in which digital money stays circulating among merchants, billers, and businesses without being cashed out have struggled to pass 20 to 25 percent of volume. Even inside Kenya, paper cash remains the default medium outside major towns and across heavy industries like farming and construction. Next door in Tanzania, where Vodacom Fintech Group Chief Commercial Officer Epimack Mbeteni spent seven years growing the business, cash still dominates primary commodity trade.

For telecoms chief executives and their shareholders, leaning so heavily on the CICO engine has become a balance-sheet risk. Charging customers a toll to deposit cash, send it to a relative, and pull it out at a kiosk used to be a high-margin business protected by physical distribution. Once an operator painted the roadside kiosks, rivals struggled to break in. Today, regulatory pressure, inflation, and instant bank-to-bank switches are compressing those withdrawal fees. Meanwhile, newer value-added products carry much tighter margins. In Kenya, Safaricom’s Fuliza overdraft facility moves trillions of shillings a year and gives millions of urban households a private way to borrow KSh 2,000 on a phone without asking neighbours for help. Yet Fuliza brings in less than 6 percent of top-line revenue because operators keep fees low to avoid the regulatory pushback that has hit transfer tariffs.

In response, Vodacom Group’s Vision 2030 strategy requires non-core financial services, spanning lending, merchant payments, wealth management, and insurance, to grow three to four times faster than basic peer-to-peer transfers, reaching at least a 50 per cent revenue share by 2030. Group-wide, those services currently bring in between 30 and 40 per cent of revenue. Tanzania has already passed the 2030 target, generating roughly 60 per cent of its M-Pesa revenue from ecosystem products and under 40 per cent from CICO tolls. Hitting that equal balance across all eight of Vodacom’s African markets means abandoning the habit of copying Nairobi’s playbook across borders.

When Vodacom Group brought its eight financial services markets under one roof (combining M-Pesa operations in Kenya, Tanzania, Mozambique, Lesotho, and the Democratic Republic of Congo with South Africa and Egypt’s Vodacash), management had to deal with wildly different stages of market maturity. Mbeteni compares running pan-African fintech to raising a family with one child in nursery school, a second in university, and a third finishing a postgraduate degree.

In a frontier market like the DRC or Ethiopia, literacy is lower, smartphones are scarce, national ID registries have big gaps, and a mobile wallet acts as the first formal account a citizen ever holds. South Africa sits at the opposite end of the spectrum. Commercial banks dominate the financial system, formal supermarket chains and card terminals handle retail commerce, and Buy Now, Pay Later credit works because credit bureau files and card rails already cover the working population.

Vodacom tried twice to launch standard East African M-Pesa in South Africa and failed both times. Outside observers were puzzled because millions of South African wage earners still pull 100 per cent of their paycheques out of ATMs on payday, worried that bank charges and automated debit orders will eat into their balances. Mbeteni points out that the East African wallet failed in South Africa because the customer’s money already started inside a formal bank account capable of digital transfers, while formal supermarket checkouts already handled over-the-counter bill payments. A standalone peer-to-peer wallet offered nothing that a South African bank account and a Shoprite till did not already provide.

“South Africa is a very different story. We tried and failed, by the way, twice we failed. And the lesson learned was that you cannot just copy and paste. In East Africa, people did not have money in the bank. In South Africa, the source of money is already the bank.”

Epimack Mbeteni, Chief Commercial Officer, Vodacom Fintech Group

A similar miscalculation tripped up the first wave of African super-apps. Borrowing from China’s WeChat and Alipay, telecoms firms and venture-backed fintechs packed dozens of lifestyle mini-apps for ride-hailing, food delivery, event tickets, and online shopping inside their payment screens. In cities like Nairobi and Johannesburg, however, smartphone users had already settled on their favourite apps. A Nairobi office worker chats on WhatsApp, orders lunch on Glovo or Uber Eats, and opens the M-Pesa app with one habit in mind, which is to send money or clear a bill and lock the phone.

Stuffing the home screen with dozens of shallow mini-apps overwhelmed users, who simply ignored them. Mbeteni admits that operators chased the sheer number of mini-apps instead of building deep utility in a few essential services. In its latest app updates across Kenya, Tanzania, and South Africa, Vodacom has stripped out the clutter to focus on a small group of high-volume mini-apps, led by municipal and government services that have never been digitized elsewhere.

For fifteen years, banks and telecoms operators treated business digitization as a matter of handing a shopkeeper a till number or QR code and counting them as an active merchant. A bare payment till does nothing to run the daily operations of a farm, a quarry, or a pharmacy. Across the continent, operators are now testing four models that pull cash out of supply chains by building software tailored to specific industries.

Vodacom is reorganising its merchant business around sector-specific platforms like Timiza na Kilimo in agriculture, where farmers have historically preferred cash to avoid unexpected digital loan deductions. Built through partnerships and targeted software acquisitions on top of Vodacom’s core ledger, Timiza na Kilimo combines bulk buyer payouts, farmer collection wallets, transaction histories that can support credit scoring, and crop insurance inside one workflow.

A well-known case from the stone quarries of Thika shows how custom workflows displace cash where generic tills fail. For years, building contractors drove to Thika carrying sacks of banknotes to buy cut stone, leaving quarry owners sitting on millions of shillings in cash every evening. Because bank branches sat kilometres away in town, armed motorcycle gangs regularly ambushed the quarry owners on the road. At one point, quarry owners found it cheaper to put the local gangs on a retainer to escort the cash to the bank than to risk daily robberies. Physical cash only disappeared from the pits once Safaricom built a dedicated Lipa na M-Pesa B2B collection workflow tailored to how lorry drivers, brokers, and quarry owners actually settle orders. Mbeteni notes that the same gap remains wide open today in Tanzania’s artisanal mining belts and across regional bus, ferry, and domestic airline corridors, where clerks still carry sacks of notes because no one has wired payments into their dispatch and ticketing software.

In Nigeria, Tayo Oviosu watched competitors spend $100 per POS terminal and give the hardware away free to chase low-margin cash withdrawals. Repeating his rule that strategy means choosing what not to do, Oviosu refused to burn Paga’s equity subsidising free CICO hardware. Instead, alongside Paga Engine (its B2B infrastructure rail serving over 300 corporations) and its retail wallet, Paga set up Doroki as a separate, non-regulated software company.

Doroki sells small supermarkets, pharmacies, and restaurants the software they need to run their shops. In a restaurant, Doroki tracks back-room food inventory, table bookings, waiter orders, kitchen ticket printing, and stock depletion, while offering working-capital loans based on sales velocity. Oviosu made Doroki open to every competitor’s payment rail. A restaurant using Doroki can take payments through OPay, Moniepoint, Access Bank, GTBank, or UBA card terminals. When rival bankers ask why Paga lets their POS machines plug into Doroki, Oviosu points to the unit economics. By running the merchant’s inventory ledger and daily software workflow, Doroki earns software subscriptions and lending margins while leaving banks to fight over fractions of a per cent on card switching fees.

After its two early M-Pesa false starts in South Africa, Vodacom stopped pushing peer-to-peer wallets and used its base of more than 30 million mobile subscribers to sell products where traditional banks were slow. Alongside partnering with Alipay on the VodaPay app to move utility bill payments off supermarket queues and onto phones, Vodacom focused on insurance and merchant cash advances. Even though South Africa has the highest insurance penetration in Africa, Vodacom found underserved segments and built one of the country’s most profitable insurance books. On the lending desk, it skipped consumer overdrafts to offer working-capital advances to small merchants against their payment turnover. In Kenya, Safaricom is making a similar move up the value chain for urban earners through its Ziidi money-market wealth product.

Across African markets, WhatsApp voice notes far outstrip text messages among informal traders because reading and typing menus in English or French shuts out millions of users. Mbeteni remembers visiting a Maasai village in northern Tanzania and travelling through the Lake Region two years ago. Assuming every Tanzanian understood Swahili, he stood up to explain Vodacom’s financial products to the village and quickly realised he needed a local interpreter. Expecting a pastoralist or smallholder farmer who speaks neither English nor standard Swahili to navigate a seven-step USSD menu tree is unrealistic. Vodacom’s Vision 2030 roadmap therefore leans on conversational voice AI that works in local mother tongues over basic feature phones, allowing a rural customer to speak a withdrawal or payment instruction in their own language for the system to verify and execute.

Whenever central bankers or members of parliament debate digital payments, a common populist demand is that operators cut transfer tariffs in half or make person-to-person transfers permanently free to wipe out cash.

Transaction records across East Africa show that pricing is rarely the main reason people stick to banknotes. In Kenya, more than 60 per cent of M-Pesa transactions already carry a zero tariff for the consumer, including small transfers of KSh 100 ($0.80) and below, and Vodacom’s average fee per transaction has fallen steadily against inflation. Across every market the group operates in, buying airtime or mobile data through M-Pesa is completely free and frequently comes with bonus airtime. Even so, in several Vodacom markets fewer than half of subscribers use M-Pesa to top up their own phones, walking instead to a roadside kiosk to buy paper scratch cards with cash.

The five-kilometre agent gap

East Africa covers a landmass nearly the size of Western Europe, with wide stretches between rural settlements. If a farmer wants to pay a village shopkeeper digitally, yet the nearest CICO agent where they can load cash into their wallet sits five kilometres down an unpaved road, no sensible person will walk ten kilometres round-trip just to make a digital payment at a corner shop.

Missing identity documents and handsets

Before a citizen can register a SIM card or open a mobile wallet to use a free payment rail, statutory KYC rules require a government-issued national ID card. Millions of rural residents across the DRC, Tanzania, and parts of Kenya still lack formal identity papers or the cash to buy an entry-level phone.

Low opportunity cost of time in underemployed towns

In markets with high formal unemployment, Western assumptions about saving time at the bank fall apart. In countries like Burundi, when bankers tell a market trader that mobile payments will save a two-hour trip to the bank branch, the trader will often reply that they have no other appointments that afternoon and enjoy walking into town to visit the branch. For subsistence households eating what they harvest and buying little more than salt and cooking oil, a digital wallet has little work to do until local farm output and trade pick up.

Because cash persistence stems from physical infrastructure gaps and compliance friction rather than transfer fees alone, FINTAK and Precursor are pressing policymakers and regulators on four practical reforms.

Accelerating grass-roots national ID registration, tiered KYC, and handset tax relief

Regulators cannot digitize citizens who lack legal identity papers or cannot afford a screen. FINTAK will engage the Ministry of Interior, the Central Bank of Kenya, and regional communications regulators to speed up grass-roots national ID registration, introduce risk-based Tier-1 KYC rules for small informal traders, and cut import duties and excise taxes on entry-level smartphones.

Replacing rigid local-server mandates with regional cloud rules

Several African central banks and data protection offices still bar financial platforms from using the public cloud, forcing operators to maintain physical on-premise servers inside every national border. While global technology firms like Meta serve 54 African countries from unified cloud infrastructure, groups like Vodacom have to duplicate costly server hardware country by country across 103 million customers. FINTAK supports regional cloud rules that let licensed institutions run encrypted, tier-one cloud architecture with real-time supervisory access for local regulators, cutting the cost of rolling out AI and credit scoring.

Opening municipal and government payment APIs to all licensed rails

Vodacom’s usage numbers show that e-government mini-apps convert cash users faster than lifestyle marketing. FINTAK is urging national ministries and county governments to publish open, non-exclusive APIs for commuter rail, ferries, bus termini, county market fees, public hospitals, and farm input subsidies so any licensed payment provider can plug in.

Enforcing POS terminal sharing and open merchant APIs

Importing millions of locked, single-acquirer $100 POS machines wastes scarce capital. Following the open-architecture approach demonstrated by Paga’s Doroki and cross-border switches like Onafriq, regulators should back POS hardware sharing and open merchant APIs so a retail software system can route payments through any licensed bank or fintech acquirer.

The first twenty years of African fintech belonged to whichever company painted the most roadside kiosks and signed up the most cash agents. As instant payment switches turn basic money transfers into a low-margin utility, physical distribution is no longer enough to defend market share.

Over the next two years, the institutions that pull ahead will be those building software for specific African trades, from the stone quarries of Thika and smallholder farms in Tanzania to independent pharmacies and restaurants in Lagos. Paper cash will lose ground when industry software makes running an African business safer, simpler, and more creditworthy than reaching for a stack of banknotes.