Despite growing adoption of artificial intelligence (AI) in key functions such as credit scoring and fraud detection, most Kenyan banks remain AI-immature —unable to fully explain how their models make decisions—raising concerns about transparency, fairness, and regulatory preparedness.
A new survey by the Central Bank of Kenya (CBK) shows that 67 percent of commercial banks, microfinance institutions, and digital credit providers are still in the early stages of AI maturity. These institutions are either merely aware of the technology or running limited pilot projects with minimal resource investment.
Using the Gartner AI Maturity Index —a framework that assesses how advanced and effective an organisation’s AI deployment is— the CBK found that 54 percent of lenders are at Level 1, marked by basic awareness and early exploratory initiatives.
An additional 13 percent are at Level 2, where institutions are running active pilots and experiments. Both levels are considered AI-immature.
Only 24 percent of lenders placed themselves at Levels 3, 4, or 5 –categorised as AI-mature—indicating moderate to advanced usage of AI across operations. The remaining nine percent have not considered AI use at all.
Credit reference bureaus (CRBs), are the least AI-mature, with all of them falling in Level 1 or 2, while commercial banks show the highest level of maturity, with 34 percent of them falling under levels 3, 4 and 5.
With such low maturity levels, only 50 percent of lenders have actively adopted AI in their operations, with credit risk assessment and fraud risk management being the top applications.
Yet, 44 percent of AI adopters admit they cannot adequately explain how their models work, according to the CBK survey. This is compounded by the fact that many institutions rely on third-party vendors.
Of the institutions using AI, 46 percent developed the applications in-house, 40 percent outsourced development, and 24 percent partnered with other entities.
“Financial institutions have adopted different approaches in the implementation and usage of AI-based applications based on their operational needs, resource availability, and strategic goals,” said CBK in the survey.
Explainability was also cited as one of the top risks in AI deployment. The CBK noted that some banks are attempting to address this by employing tools and techniques to interpret AI decisions and enabling human review and intervention where necessary.
As AI use in Kenya’s financial sector deepens, the findings point to a critical gap between adoption and readiness —underscoring the need for improved governance, internal capacity, and responsible use frameworks.
