AI for Inclusion: Rethinking Credit Scoring in Data-Poor Environments
Mamari Keita- Principal Consultant, Business Solutions, Temenos
Across much of the world, people are still being excluded from the formal financial system. This is not because they are unbankable, but rather due to the system not seeing them. These are the “credit invisibles”: people who do not have formal payslips, credit histories, or other types of structured data typically used by banks to assess lending risk. Many live in cash-based or informal economies, especially across Africa, Latin America, and parts of Asia. Others are self-employed, undocumented, or simply off the grid. They are not necessarily risky, just invisible.
At Temenos, we believe that it is a solvable problem. With responsible use of artificial intelligence (AI) and access to alternative data, we can start building better, fairer models that do not punish people for where they come from or what data they lack, but that include them on their own terms.
What makes a system inclusive?
We often talk about inclusion in broad strokes. For instance, more access, reach, and fairness. But when it comes to credit scoring in particular, inclusion starts with a basic idea that if someone does not show up in your data, that does not mean they do not exist.
Traditional scoring models assume a level playing field, such as stable employment, predictable income, access to utilities, and historical repayment patterns. But what happens when none of that applies?
In so-called data-poor environments, that assumption falls apart. The inputs become misaligned. And so, too often, the result is exclusion.
A new lens on risk
What is needed is a shift in mindset. We need to ask what other signals we can responsibly use to assess creditworthiness.
At Temenos, we have worked with financial institutions in diverse markets that are asking exactly that. Whether it is analyzing mobile money usage, digital behavior, or even local community reputation scores, we are helping banks think beyond the traditional data playbook.
With our AI-powered Explainable Decisioning engine, we have built in flexibility to support both conventional and alternative data sources. And because we prioritize explainability, institutions can open their risk models, seeing what factors influenced the score, and why.
That transparency matters, especially when experimenting with new forms of data. It gives banks the confidence to innovate, while keeping regulators, boards, and customers on side.
Fairness is not optional
Inclusion without fairness is not inclusion at all. As we expand the data inputs into AI models, we must be even more vigilant about unintentional bias, data privacy, and consent.
That is why we have designed our AI tools to be ethical by default. Every decision can be explained, every dataset can be audited, and every model can be tuned to reduce bias, not reinforce it.
We also believe in putting people in the loop. Banks must maintain human oversight of AI-driven decisions, especially when it comes to sensitive outcomes like loan approvals or denials. Our tools enable banks to set up guardrails and override mechanisms, so that automation never becomes a black box.
Reaching the credit invisible
We have also seen this work in practice. In one project with a client in Africa, in a country where no credit bureau information is available, we had to use alternative information in order to build accurate models. And because our modelling approach is flexible, we can either leverage conventional data types or alternative data types such as mobile money data, top-ups, transfers, and bill payments, as a proxy for financial responsibility, when available. We can deploy data-based models or rules-based models when no data is available at all.
That one shift made it possible to extend credit to customers who had never had a formal bank account. And because the model was explainable and tested for bias, the bank could justify every decision, internally and externally.
This kind of innovation is not about taking shortcuts or lowering risk thresholds. It comes down to bringing more people into the system on fair, transparent, and ethical terms.
A glimpse into the future
I am optimistic for the future. Over the next few years, I believe we will see AI open new doors to inclusion in the everyday lives of people who have been excluded for too long. As data becomes more accessible and digital infrastructure expands, banks will have more tools than ever to build services that are truly inclusive.
But we need to stay grounded. Inclusion is not something that happens automatically when you introduce new tech. It is not just a question of algorithms but rather about how you design systems, how you govern them, and how you make decisions that reflect real human needs.
That is why, at Temenos, we are working side by side with banks to make inclusion intentional. It should not be a by-product of innovation but the goal from the very beginning.
Let’s connect
Resources
Digital Banking
Quickly deliver innovative, AI-powered financial services with Temenos Digital Banking, a cloud-native, flexible platform.
Temenos for Inclusive and Community Banking
A scalable, world-class solution bringing ultimate efficiency to banking to enable financial inclusion around the world.
Leveraging Technology to Transform Community Banking
Explore how digital innovation is transforming community banking and driving financial inclusion across MEA’s underserved regions.