Treasury departments around the world are discovering hard truths. In the digital age, if government IT is clumsy, government itself is clumsily slow, expensive, and frustrating for citizens and businesses alike.
The African Cooperation Budget Reform Initiative (CABRI) Regional Peer Learning and Exchange Workshop held in Kigali, Rwanda this month held ODI Ameya Ashok Naik, Development and Finance Research Fellow at Global, said the latest emerging themes from work from ODI’s Digital Finance Hub demonstrate how far we have come in digitalizing finance, and how far we still have to go.
On paper, the Treasury looks surprisingly modern. Most companies now use digital systems for core functions such as budgeting, financial operations, and financial reporting. But beneath these sophisticated dashboards lies the troubling reality of fragmented platforms that barely communicate with each other. Interoperability between financial management information systems (FMIS), e-procurement systems, and public investment systems is partial at best, making it extremely difficult to combine financial and non-financial data for real-time decision making. The result is a hybrid environment where parts of the cloud and APIs are bolted onto paper files and legacy software, where data validation and a shared “source of truth” are the central battle, rather than plumbing in the background.
Then comes AI, the buzzword of the decade. Treasury departments are under pressure to deploy artificial intelligence in everything from revenue forecasting to fraud detection. But this report is quite sobering. For the most part, AI in finance is still in the experimental stage, an extension of existing analytics rather than a clean revolution. Although predictive models can outperform traditional methods, their inner workings are often opaque, raising deep questions about their explainability, legality, and reliability in budgeting and auditing processes.
This leads to a major strategic dilemma. Should the Treasury pour scarce resources into ambitious AI pilots, or invest in people and skills that can challenge sold-out vendor promises and protect the nation’s data assets? ODI suggests the second path could be just as important as the first. Without skilled internal teams, governments risk vendor lock-in, duplicate systems, and low-value contracts that secretly mortgage future fiscal space.
To overcome this situation, this paper argues for a shift to total cost of ownership thinking. Rather than viewing digital systems as a one-time capital project, agencies need a lifetime model of costs and benefits that covers development, maintenance, upgrades, integration, and eventual replacement. This requires new capabilities, including multidisciplinary teams, agile ways of working, user-centered design, and funding models that recognize software as living systems rather than concrete bridges.
The message is clear. Digital finance is no longer about “buying an FMIS.” It’s about building a nation that understands its own data, controls its own algorithms, and treats technology choices as long-term financial decisions. Agencies that acquire these skills will not only procure better systems, they will also quietly rewrite the face of effective and responsible government in the 21st century.
Dr. Yusuf Musherle is a Senior Lecturer and Consultant
