For many years, the main measure of financial inclusion was account ownership: how many adults have an account at a bank or with a mobile money provider. It is a sensible starting point. You cannot use a service you do not have. But the World Bank's Global Findex surveys have also shown, repeatedly, that a meaningful share of accounts sit inactive. Access is necessary. It is not the same as inclusion.

If people have an account and still choose cash, the informal savings group or a trusted shopkeeper, that choice is information. It usually tells us that the formal option does not fit their lives as well as the alternatives do. That makes inclusion, in large part, a design problem.

People already have financial lives

It is easy to describe the unbanked as people without financial services. In practice they manage money constantly: rotating savings groups, informal credit from suppliers, money kept with relatives, careful timing of purchases around irregular income. These arrangements are often well suited to their circumstances. They are flexible, built on personal trust and available without paperwork.

A new financial product does not enter an empty space. It competes with these arrangements, and it has to be better on the terms that matter to the user, not the terms that matter to the provider.

What good design looks like

Products that succeed with people new to formal finance tend to share a few qualities.

  • Clarity about cost. Fees are stated up front, in plain language and in the local currency, before the user commits.
  • Room for irregular income. Savings and repayment schedules that bend with uneven earnings rather than penalising them.
  • Recovery from mistakes. A clear way to reverse an error or reach a person, because one lost transfer can end someone's trust in digital money.
  • Respect for language and literacy. Interfaces that work in the languages people speak and do not assume fluency with forms.
  • Resilience to poor connectivity. Services that degrade gracefully when the network is slow, instead of failing at the moment of payment.

Where AI can help, carefully

Artificial intelligence can widen access in real ways: assessing creditworthiness for people without formal credit histories, detecting fraud that targets new users, and offering support in local languages at a scale human teams cannot match. Each of these benefits comes with a responsibility. Models trained on narrow data can exclude the very people inclusion efforts aim to reach, and automated decisions about money need to be explainable to the person affected.

The test for any of these tools should be the same as for any other design choice: does it make the service fit people's lives better, and can they trust it?

Measuring what matters

If inclusion is a design problem, then success should be measured the way designers measure it: by use, not by sign-ups. Are people still using the service after six months? Do they trust it with larger amounts over time? Do they recommend it to family? Those questions are harder to answer than counting accounts, and far more honest.

The opportunity across African consumer markets is not only to open more accounts. It is to build financial services that people choose, because they are clearer, fairer and more reliable than the alternatives they already know.