AI in Government

Public financial management (PFM) has long embraced new technologies — from financial information systems to data analytics and robotic process automation. AI represents the next phase of this evolution, helping governments enhance forecasting, improve oversight, and automate routine processes. In most cases, AI is not transforming PFM from the ground up, but rather reinforcing and accelerating the efficiency of existing systems. As such, it acts more as an assistant or advisor than as a replacement for human decision-making.

The current state of play

PFM applications of AI remain focused on enhancing — rather than replacing — traditional processes. Five core areas of use are emerging:

• Improving forecasting. AI improves the accuracy and timeliness of economic forecasts and enables nowcasting of indicators like GDP and inflation. Governments are experimenting with models that outperform traditional approaches while aiming to preserve interpretability.
• Facilitating spending decisions. AI can analyse historical and real-time data to detect spending patterns, flag inefficiencies and evaluate the effectiveness of public programmes. These insights support better allocation of public funds.
• Supporting budget planning and monitoring. By modelling baseline costs and future spending needs, AI supports more accurate budgeting. It also enables earlier detection of fiscal risks and performance issues using predictive analytics.
• Automating management, reporting and oversight activities. Natural language processing and machine learning help automate repetitive tasks such as document classification, invoice matching, and internal controls — enhancing both speed and accuracy in financial oversight.
• Facilitating engagement with stakeholders and users. AI-powered chatbots and digital assistants are increasingly used to guide users through financial services, fiscal data, and benefit systems. These tools improve accessibility and responsiveness in financial governance.

Despite this progress, many finance ministries face legacy IT systems, fragmented data, legal uncertainty, and skills shortages that limit AI adoption. Ensuring transparency and maintaining human oversight remain essential for trust and accountability.

Examples from practice

• Sweden: Transparent forecasting with explainable AI. Sweden’s National Financial Management Authority uses explainable machine learning to improve GDP forecasts and make their underlying logic visible — supporting more accountable macroeconomic planning.
• Korea: Real-time fiscal decision support. Korea’s dBrain+ system integrates AI across budget, risk, and performance functions — linking 63 systems and agencies to enhance real-time financial analysis and coordination.
• France: Early warnings for municipal finance risks. France’s tax agency developed an AI system to identify municipalities at risk of financial distress, enabling early intervention. The tool uses clustering and historical data to distinguish between temporary and structural issues.
• Brazil: Enhancing fiscal transparency. Brazil’s National Treasury uses AI to classify government spending and has reduced processing time from 1,000 hours to 8. The initiative has expanded to include climate-related expenditures, improving fiscal accountability.
• United Arab Emirates: Fiscal chatbots for citizens. The UAE’s U-Ask chatbot provides automated answers to common fiscal queries, including on reporting and budget matters — offering a unified, user-friendly interface across government.

Untapped potential and the way forward

Finance ministries have focused on predictive AI, but prescriptive AI — which recommends actions — could reshape core public financial management functions. It may shift accountability, oversight and policy design. Governments will need evaluation frameworks to track impact, cost and fairness; alongside promoting transparency, engagement and upskilling. Starting small can build trust and capability for broader adoption through a coordinated, learning-oriented approach.

Learn more

Review a detailed section on AI in public financial management here.