How data and AI are reshaping access to finance

Access to financial services like banking and credit is a foundational driver of inclusive and sustainable economic development. By expanding access to financial services, enabling innovation, and facilitating people to buy, sell, and compete in the market, households and businesses can be empowered to grow, and economic opportunities are unlocked.

Over the past decade, J-PAL and its research network have built a substantial body of evidence on how to effectively improve people’s—and businesses’—access to digital financial services, credit, and other dimensions of financial inclusion. 

In 2025, J-PAL Finance sector Co-Chairs Emily Breza and Emanuele Colonnelli convened a group of researchers in Nairobi, Kenya with J-PAL Affiliated Professor Christopher Woodruff and Private Enterprise Development in Low-Income Countries (PEDL) to engage with finance-sector stakeholders, including representatives of banks, lending groups, and financial investors to advance the research agenda on inclusive finance for development. These discussions highlighted five priority areas for financial inclusion in Africa: 

1.  How data and AI are reshaping access to finance,
2. Building venture markets where none exist: Coordination, capital, and evidence,
3. Where international finance meets development: The role of currency risk,
4. Why people matter in high-growth entrepreneurship in Africa, and
5. Financing challenges along agricultural value chains.

This is the first post in a new blog series that will explore each of these themes in turn, examining their role in shaping financial systems, the frontiers of technological innovation, and key evidence gaps that call for further research. Today, we’ll focus on how data and AI are reshaping access to finance. 

From mobile money to AI‑driven finance

Many African countries have expanded financial inclusion rapidly through mobile money and digital payment systems. Since Kenya has been at the forefront of this global expansion, several conversations in Nairobi focused on how a new wave of AI-driven tools builds on this progress, and has the potential to further expand access to finance, reduce costs, and tailor services to people's needs.

These tools use AI in four areas: revenue-based financing, consumer credit, insurance, and cross-border payments. There are important research gaps in these areas, which matter especially in Africa, where AI infrastructure is still emerging, and policymakers are looking for concrete evidence to shape new AI strategies.

AI across financial services in practice: Credit, insurance, and payments

Traditional financing models in African markets often exclude SMEs that lack conventional collateral or credit histories. Revenue-based financing offers an innovative alternative by providing capital to businesses (firms) based on ongoing revenues, rather than fixed assets. Instead of fixed monthly payments, firms repay a share of their revenues, so payments can rise and fall with business performance. This can make financing more manageable during slow periods and reduces the risk of default. 

At the same time, evidence shows that there are contracting challenges: borrowers may shift parts of their sales elsewhere to slow the pace of repayment, or firms that expect their incomes to fall may be disproportionately attracted to this type of financing. However, equipment or digital systems that automatically record usage and sales can prevent borrowers from shifting sales elsewhere and ensure more accurate reporting. 

In Nairobi, one of the examples of this approach comes from Untapped Global. Untapped Global finances high-growth firms by monitoring how their equipment is actually used and how much revenue they generate in real time. This gives investors transparency and confidence and allows entrepreneurs to access capital without needing traditional collateral. Companies like Fleetsimplify make this possible for transportation businesses by using AI-based tools to track vehicle use and ride-hailing earnings, and then sharing this data with Untapped Global to adjust loan terms based on these real-time business performance metrics. 

Kenya’s leadership in digital payments, driven by M-Pesa, has also created a powerful source of data for lending and consumer credit. For many households and small businesses that do not have payslips or formal accounts, mobile-money histories offer a clear picture of their day-to-day cash flows. Lenders are already using these digital trails to assess creditworthiness, including with the help of AI-tools that spot spending and repayment patterns and make sense of unstructured transaction descriptions.

Fintech lenders like 4C Group show how these tools can help match loan sizes to what small firms can realistically repay. Patascore, a credit-scoring firm in Nairobi, highlighted the next step: agentic AI systems that can adjust loan terms, borrowing limits, and prices almost instantly as customers’ circumstances change. 

Yet, despite the availability of data, many lenders still offer the same loan product to everyone. That leaves significant room for products that respond to people’s actual needs and risks.  For example, evidence shows that microequity, i.e. small, share‑based investments, can give borrowers more flexibility than traditional loans or debt contracts, and ongoing work demonstrates the potential of using experiments and machine learning to set loan prices that better match how different borrowers behave. 

AI is also reshaping how insurance works. ZEP-RE, a pan-African reinsurer based in Nairobi, is using AI to make livestock insurance more reliable in pastoral communities, where animals often double as income and collateral. Because fraud can occur where insurers cannot differentiate between animals, ZEP-RE is using AI-powered muzzle-recognition—a sort of “facial ID” for cattle—to verify an animal's identity. This helps build trust and increases the value of livestock as collateral. 

AI is also helping ZEP-RE price insurance more accurately. By incorporating weather forecasts, they can better estimate crop and livestock risks—especially when payouts depend on fodder availability, an early and easily measured signal of drought‑related losses. The same forecasts are used to help nomadic herders decide where to move their livestock and to offer support, such as fodder, to those who stay in high‑drought-risk areas.

Beyond agriculture, AI shows potential to improve how insurance claims are handled. Claims processing is often slow for policyholders, which reduces the real value of having insurance. ZEP-RE is developing AI tools to process unstructured health-insurance claims so that approvals and payments happen faster. AI may also help detect potentially fraudulent claims, a common challenge in health insurance, potentially lowering costs for insurers and customers alike.

AI is already changing how cross‑border payments work, especially when it comes to managing currency risk. As we’ll discuss in the third post of this series, a key challenge in sending money across borders is navigating foreign exchange risk. New AI tools are helping reduce this uncertainty. 

For example, the venture capital firm Flourish Ventures highlighted how AI agents are being used to search for the best available exchange rate. Onafriq, a pan-African payments company, uses AI models to predict short-term exchange-rate movements that take place between transaction initiation and settlement. By reducing surprise fees and hidden exchange‑rate charges, these tools could make cross-border payments cheaper and more predictable - especially important to remittance recipients and small businesses, where even small differences in cost can affect a family’s budget or a firm’s ability to operate.

Evidence gaps and opportunities

Despite rapid experimentation, many AI‑driven financial innovations in Africa remain at an early stage, with promising ideas but limited large‑scale implementation. Progress is slowed by practical challenges: limited local data‑center capacity, unreliable electricity, shortages in AI‑related skills (stay tuned for our forthcoming blog post on why people matter in high-growth entrepreneurship), and the lack of large language models trained on African languages. 

Firms like Qhala in Kenya are actively working to close some of these gaps through advocacy and digital transformation projects. At the same time, policymakers are building data governance, consumer protection, and privacy frameworks, while the underlying technology continues to evolve.

In addition, AI systems learn from past data, which can reinforce historical biases and existing inequalities. As a result, people with thin digital footprints or past financial difficulties risk being excluded, and it is often unclear what models are actually learning from these datasets. The challenge is therefore not only about better technology, but about designing AI-enabled financial services that people can understand, trust, and actually benefit from. 

Better understanding and addressing the issues outlined above requires stronger evidence—in other words, more rigorous research. This research can help inform responsible product design, guide policy choices, and ensure that the next wave of AI‑enabled financial services expands access rather than deepens existing gaps. 

The examples discussed above point to four open questions where more research is especially needed:

1. What are the effects of revenue-based financing?

The examples above highlight the promise of revenue-based finance; while research exists on flexible repayment structures of micro-finance contracts, and evidence on comparing asset based financing under a traditional debt contract to performance-contingent microfinance contracts shows large positive impacts from the contractual innovations, several questions remain to be better understood:

How does revenue‑based financing affect firm growth trajectories, profitability, and financial resilience over time, compared to traditional debt or equity financing in African markets? Which types of firms benefit most from revenue‑based financing? How does the use of real‑time data and automated monitoring affect access to finance? What are the implications of revenue‑based financing for financial inclusion, especially for asset‑heavy or data‑rich SMEs in African markets?

2. How does the use of mobile money data and AI‑driven credit tools affect access to credit, loan terms, and financial outcomes for households and small businesses?

Can improved credit product customization unlock higher return investments for small firms? Do AI‑based credit models meaningfully expand access to credit for borrowers with limited formal histories, or mainly reallocate credit among existing users? What constraints limit adoption of risk‑based pricing? Under what conditions can agentic AI systems improve both financial inclusion and lender sustainability?

Answering these questions is key to understanding whether AI‑enabled lending becomes a tool for broader inclusion—or primarily a refinement of credit for those already served.

3. How can AI‑enabled insurance tools improve access, affordability, and trust in insurance, particularly for underserved populations?

When do AI-based tools genuinely expand insurance coverage, and for whom? Can AI-enabled insurance applications raise the customer experience and lower costs? How do improvements in pricing accuracy, fraud detection, and claims speed affect customer trust and long‑term participation in insurance markets? Under what conditions do AI tools reduce costs enough to expand coverage at scale?

4. How does the use of AI in cross‑border payments affect costs, pricing transparency, and trust for households and businesses?

Do AI‑driven tools meaningfully reduce the total cost of cross‑border payments? Who benefits most from these cost reductions - consumers, payment providers, or intermediaries? How do increasingly autonomous, “agentic” AI systems change pricing, competition, and transparency in consumer‑facing payment markets? Do these technologies increase usage among households and small businesses, or mainly improve efficiency within existing payment channels?

J-PAL’s Finance policy team is working to support new research projects that answer important questions on financial inclusion and innovation and draw out results from multiple studies to identify key insights relevant to policymakers and financial institutions. Stay tuned for future posts in this series outlining more evidence-informed approaches to expanding access to financial services, enabling innovation, and building economic empowerment. Subscribe to receive updates.

We would like to thank the finance stakeholders we met in Nairobi for their insights and openness: Antler East Africa, Baridi, Beyond Capital Ventures, British International Investments (BII) in Kenya, Central Bank of Kenya, Enza Capital, Equity Bank Kenya, FASA - Financing for Agricultural SMEs in Africa Fund, Fleetsimplify, Flourish Ventures, IETP - Investisseurs et Partenaires, IFC Kenya, Ketha Africa, Kukua, NALA, NCBA Bank Kenya Plc, Norfund, Onafriq, Proparco, Pure Infrastructure Ltd, Qhala, Sayuni Capital, Stanford Seed East Africa, TLCom Capital, TLG Capital, Untapped Global, VestedWorld, ZEP-RE (PTA Reinsurance Company), and 4C Group.