What survived the predictions about African fintech?

In 2021, at the height of the global zero-interest-rate venture boom, industry forecasters sketched an ambitious five-year map for African financial services. The consensus predicted a wave of shared industry platforms, open-data aggregation, rapid institutional digitisation, blockchain-based registries and widespread adoption of digital fiat and central bank digital currencies across cross-border trade, all overseen by an increasingly assertive regulatory state. The forecasts were not foolish; they were simply priced for a world of free money, and that world ended.

Looking back from November 2026 across the wreckage and the triumphs reveals a sharp lesson for every chief executive, investor and central banker drafting a 2027 strategy. Technology forecasts almost always overestimate the speed of shiny new rails while underestimating the stubborn institutional gravity of governance, unit economics and household cash flow. In June 2023, after canvassing responses to its February 2022 discussion paper, the Central Bank of Kenya quietly shelved plans for a retail shilling digital currency, concluding that the existing mobile-money and bank ecosystem already solved retail speed, and that a state digital currency would introduce bank-disintermediation and cybersecurity risks without fixing a clear consumer pain point.

By contrast, where sovereign digital infrastructure solved a genuine distribution bottleneck, the central bank's DhowCSD portal letting everyday savers bid directly on Treasury bills from a phone, or mobile money-market funds scaling retail wealth management, adoption took off, because the tool matched an immediate economic need. The predictions that failed forgot to ask which institution would pay to keep the pipe running.

Counting live operating transactions instead of consortium press releases.

Take the 2021 prediction that smaller African banks, microfinance lenders and SACCOs would rapidly pool their technology stacks onto shared utility platforms. On paper, hundreds of memorandums of understanding were signed over the past five years. On the ground, many shared-service dreams stalled over three governance questions no press release ever answers: who pays the fixed cloud bill when two members drop out, who owns the customer relationship when a shared scoring model cross-sells a product, and who carries the regulatory fine when the shared vendor goes offline.

As the Financial Stability Board warned in its October 2025 monitoring report on artificial intelligence in finance, concentrating multiple institutions onto a handful of third-party model and cloud providers turns micro-level cost savings into macro-level systemic concentration risk. A credible forecast for shared platforms in 2027 cannot count signed partnership announcements or hackathon pilots. It has to count live, paying institutions running core production workloads with verified data isolation and tested exit chutes.

Measuring who captures the margin from data aggregation and digital money.

The 2021 expectation that open APIs and alternative-data aggregation would automatically democratise credit ran into a similar institutional wall. Kenya's September 2026 draft National Payments Policy and Bill finally laid out a statutory blueprint for a national instant payment switch, mandatory open API standards and standalone licences for payment-initiation and account-information providers. But five years of unilateral screen-scraping and mobile-statement scoring taught lenders that more data rows do not automatically produce better credit decisions. Without basic accounting reconciliation to separate borrowed working capital and self-transfers from true business margin, and without enforceable portability contracts that stop external scoring vendors locking up the bank's decision history, buying more API feeds just adds another recurring technology fee onto a stagnant loan book.

The same commercial realism reshaped digital money. While retail central bank digital currencies stalled in Nairobi and struggled for everyday traction in Lagos, private dollar stablecoins and regulated on-chain settlement rails, brought under formal supervision by Kenya's July 2026 Virtual Asset Service Provider regulations, found immediate product-market fit among importers, exporters and freelancers hedging currency volatility and bypassing slow correspondent wires. The pattern generalises: a digital token survives when it settles an invoice faster and cheaper than a SWIFT message, backed by liquid local on-ramps, anti-money-laundering compliance and legal recourse when a transfer stalls. Monetary technology is adopted for cash-flow reasons, not ideological ones.

Following formalisation into the tax ledger and the household wallet.

Perhaps the biggest blind spot in 2021's optimism was assuming that digitising a transaction automatically improves the financial health of the business or household behind the screen. Two hard anchors in Kenya's 2026 market show why access and health are different numbers.

On the enterprise side, digital trails have become the tax collector's primary enforcement arm. From January 2026, the Kenya Revenue Authority began systematically validating business income and deductible expenses against electronic Tax Invoice Management System receipts, withholding certificates and customs declarations. For a finance director or SME owner, software that fires off fast mobile payments while failing to link the purchase order, the eTIMS invoice, the withholding deduction and the bank ledger creates a severe tax-disallowance liability at year-end. The winning enterprise fintech tools of 2027 will be unglamorous deterministic reconciliation engines that keep a trader tax-compliant without hiring a full-time accountant.

On the household side, the 2024 FinAccess survey by the Central Bank of Kenya, the national statistics office and FSD Kenya documented that formal financial access climbed to 84.8 per cent of adults, up from 26.7 per cent in 2006, while the share classified as financially healthy sat at just 18.3 per cent, less than half the 39.4 per cent recorded in 2016. Four out of five adults hold an account; more than four out of five cannot absorb a shock without distress. When inflation, statutory payroll deductions and an 83.8 per cent informal-workforce reality squeeze net incomes, shipping faster 30-day overdrafts or bolting a chatbot onto a high-cost credit product merely accelerates the speed at which families borrow from tomorrow to pay for today. Better finance often looks like fewer emergency loans, fewer penalty rollovers and fewer repeat calls to a helpdesk.

Making the 2027 technology budget answerable to a date and a ledger.

As boards lock in 2027 allocations, funding vague digital-transformation envelopes has to end. Every technology proposal must first beat the cheapest non-AI operational fix on a full-cost ledger that includes human verification, error redress, regulatory compliance and the price of walking away from the supplier, while separating deterministic rules for known reconciliation from statistical scoring models, source-locked language assistants and strictly capped automated scripts.

On 31 March 2027, at the close of the first quarter, the chief executive, chief financial officer and chief risk officer should bring every pilot and platform partnership back to the board table with a named owner, an explicit unit-economics threshold and a binding kill-or-scale date. Auditing the net landed cost per completed customer outcome after human supervision, the verified cash retained by the customer after fees and the live tested cost of switching off the vendor will show which tools close the operating account cleanly, and which should be unplugged before a fashionable experiment hardens into permanent overhead. The next five-year forecast will be wrong too. The institutions that thrive under it will be the ones that budgeted for being wrong.