How AI can restructure public funds?

Governments around the world face financial pressures: rising debt, volatile income, and rising public expectations.

In this complexity, they are expected to act faster, spend smarter, and increase trust in public institutions. Artificial intelligence (AI) presents a generation of opportunities to redefine how public resources are planned, allocated and explained. However, its potential in its finances remains largely undeveloped.

The government is beginning to integrate AI into its fiscal business, from optimizing budgets and improving forecasts to automating audits and fraud detection. These early efforts suggest a bigger prize. The strategic use of AI is to redesign the fiscal policy cycle itself.

The question is not whether AI is useful, but how well it can be scaled down at a speed where the government can responsibly reduce its use.

The role of AI in modern finances

AI is transforming how governments manage public finances. Strengthen decision-making in fiscal policy and resource allocation, enhance risk management, streamline operations, and improve citizen services.

By moving beyond Simple Automation, AI enables real-time data analysis, dynamic resource targeting, and aggressive risk identification.

These features have already been applied to public finance for support.

• Macroeconomic and Financial Forecast: AI uses machine learning (ML) and deep learning (DL) to transform traditional econometric methods by processing vast, unstructured data sets. This improves prediction accuracy and enables real-time “NowCasting.” For example, Australia's tax office uses the ML model to predict tax revenues, while South Korea's Ministry of Economy and Finance uses AI to generate daily updates on government bond balances. In the United Arab Emirates, the Ministry of Finance is strengthening revenue forecasts and compliance through AI, while initiatives such as Smart Dubai have embedded intelligent tools into services such as digital payments and smart procurement.
• Budget planning and spending monitoring: AI modernizes the budgeting process by automating data processing and application of advanced analytics. ML increases the accuracy of spending baselines, supports policy cost estimation, and enables evidence-based financial decisions. For example, the Australian Veterans Affairs Office uses forecast models to simulate lifelong financial impacts of beneficiaries and evaluate policy options, while the French DGFIP applies ML to identify municipalities with financial risk, evolving from historical data analysis to predictive forecasting.
• Public spending review: AI is enhancing the spending review process by analyzing large and complex datasets to identify trends, assessing program effectiveness, and notifying resource reallocation. ML and DL extend beyond traditional analytics to uncover deeper insights and automate recommendations. For example, the UK Treasury hires HMT-GPT to assess budget proposals and support long-term funding reviews, while the Canadian Treasury uses AI to assess the impact of public spending and guides reallocation decisions.
• Accounting, Control, and Fraud Detection: AI-driven automation and anomaly detection make internal financial management more efficient. Tools using NLP, ML, and DL can quickly process documents, identify irregularities, and enhance monitoring. Denmark employs AI to monitor grant payments, flag anomalies, and the UK applies ML to detect fraudulent benefits claims with increased speed and accuracy.
• Citizen engagement and service provision: AI is redefineing how public financial institutions interact with citizens. Chatbots and language models improve accessibility, automate responses, and increase transparency. The U.S. Internal Revenue Service uses AI-powered voice and chatbots to reduce the waiting time for inquiries, while the Irish Treasury Department uses AI to draft tax manuals, summarise legal documents, and make government communication more accessible.

These examples not only highlight the role of AI across the fiscal value chain, but also reveal gaps. The AI has not informed decisions and created them.

Why normative AI remains elusive

Normative AI – the ability to recommend or make decisions – remains rare in finances.

The reasons are complicated. Lack of explanation, unclear accountability, and unresolved ethical concerns. Should the AI system determine how public funds are distributed, or which programs face cuts? What if that recommendation reflects biased or defective assumptions? Who is accountable when things go wrong? These are not just technical questions, they are governance questions. Addressing them is key to unlocking AI's next frontier in fiscal policy making.

What is suppressing AI?

Despite that promise, adoption of AI in the finances faces five lasting barriers.

• Lack of strategic alignment with institutional priorities: Many institutions lack a top-down, structured approach to identifying AI use cases that directly support state priorities or institutional obligations. This leads to fragmented, opportunistic, or siloed implementations, particularly when impact tracking is focused on cost or operational efficiency, limiting our ability to demonstrate strategic value.
• Obsolete infrastructure and fragmented data ecosystems: Legacy IT systems and non-integrated data sources hinder effective AI model development. High quality interoperable data is essential, but in many cases, you are unable or confined to a bureaucratic system that is resistant to integration. These challenges are particularly severe in areas where interagency coordination remains limited. At GCC, efforts to integrate public financial platforms, often led by sovereign wealth funds or centralized financial ministries, highlight the growing need for shared standards and interoperable systems.
• Capacity and Culture gap: AI deployment requires more than technical expertise. It demands a culture that embraces innovation and adaptive decision-making. Many agencies lack digital capabilities, operate within risk aversion environments that face internal resistance to change, or where experimentation is recommended. Regional actors such as the Arab Monetary Fund have highlighted the need for stronger institutional coordination and innovation ecosystems to promote digital finance transformation across the Arab region.
• Ethics, security, and transparency concerns: As AI begins to shape sensitive financial decisions, such as profit allocation and funding reallocation, issues of fairness, legality, and accountability become important. Without clear rules, AI can produce biased results and violate financial regulations. Weak cybersecurity can expose sensitive financial data, threaten national security and erode trust. Financial experts and citizens need to understand how AI insights are generated and used. Opaque algorithms or poorly communicated logic risks undermine both legitimacy and public trust.
• Robust assessment and lack of ROI framework: AI returns are difficult to measure and intangible in the short term. This makes it difficult to prioritize and scale promising pilots. Without a clear methodology to assess impacts, including increased efficiency, improved accuracy and equity outcomes, AI programs struggle to ensure sustained funding and political support.

Strategic paths to advance

To move from pilots to purpose-driven AI adoption, governments need to focus on five priorities:

• Anchor AI in fiscal strategy: Defines AI priorities that are aligned with institutional and national fiscal and development goals. It focuses on areas where AI is increasingly responsible for revenue mobilization, spending efficiency, and compliance.
• Infrastructure and People Investment: Build modern cloud infrastructure, hire skilled data engineers and provide ongoing training to unlock the full value of AI beyond isolated pilots.
• Enhanced data governance: Establish a strong data governance framework to improve data accessibility, quality, and interoperability while protecting privacy and promoting ethical use.
• Measure what's important: Track cost benefits metrics along with accuracy, compliance, equity, and public confidence to gain the true impact of AI on financial management.
• Embedded safeguards: Before AI tools can influence high-stakes financial decisions, model transparency, independent auditing, and a clear accountability framework are required.

Read: AI Order: 5 Steps to Transform Public Sector Services

The time to act is now

AI is not just a technical upgrade, it is a fundamental change in finance management. The government that strategically embeds it unlocks unprecedented agility, accuracy and transparency. Beyond the role of advisory, normative models can promote smarter, faster, and more accountable policy decisions.

However, this power needs to be careful. Without strict transparency, fairness and accountability protection, risk bias and unintended consequences that could undermine trust in such systems. When done responsibly, AI will help governments anticipate shocks, improve policy outcomes, and increase public trust.

For MENA countries, where fiscal reform and economic diversification are top priorities, interests are even higher. With proper investment, governance and institutional commitment, the region can not only catch up, but also lead in shaping the future of its finances.

Naman Sharma and Pedro Marques are partners, and Rayane Dandache is manager of Kearney Middle East & Africe – Financial Services Practice.