## Introduction2

## Source details

**Canonical URL:** [Introduction2](https://www.imf.org/-/media/files/publications/tnm/2026/english/tnmea2026006.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/tnm/2026/english/tnmea2026006.pdf.md)
- [Structured JSON version](/-/media/files/publications/tnm/2026/english/tnmea2026006.pdf.json)

---

### I. Envisioning Digital Transformation of PFM
- Vision: a digital PFM ecosystem built on cutting-edge digital solutions enabling real-time, data-driven decision making to achieve core PFM objectives:
  - (1) fiscal discipline;
  - (2) allocative efficiency;
  - (3) operational efficiency; and
  - (4) transparency and accountability.
- Potential benefits of emerging digital technologies:
  - increase efficiency and effectiveness by automating routine tasks;
  - facilitate data-driven decision making across the budget cycle;
  - ensure fiscal sustainability via better data analytics for fiscal policymaking;
  - strengthen transparency and accountability through stakeholder monitoring of public funds;
  - improve public service delivery via seamless, real-time interactions; and
  - promote compliance with global norms through adoption of international standards and practices.
- Illustrative "digital by design" redesign examples:
  - Budget preparation: AI agents automating data validation, reporting, and workflow approvals.
  - Public investment management: building information modeling, light detection and ranging, and AI-driven analytics.
  - Government payments: new payment systems and digital money innovations for payroll and subsidies.
  - Government contracts: smart contracts and IoT-enabled asset registries.
  - Public debt management: tokenization, distributed ledger technologies, and central bank digital currencies.
  - Performance budgeting: machine learning for target analysis and generative AI for tailored reports.
  - Internal control and audit: AI and analytics to detect anomalies and facilitate proactive, adaptive, real-time risk management.

### A. Current Digital PFM Landscape — coverage, shortcomings, and resources
- Coverage and systems:
  - Almost all countries use automated systems supporting PFM functions, typically an FMIS for budget planning, budget execution, treasury operations, cash and debt management, public investment management, and accounting and reporting.
  - Additional functions—tax collection, debt management, payroll management, and procurement—are generally automated in separate, stand-alone systems.
- Key shortcomings of many existing PFM digital solutions:
  - (1) poor quality or incomplete, inconsistent data;
  - (2) silos or lack of interoperability across databases, systems, and institutions;
  - (3) outdated IT architecture with limited flexibility, scalability, and security;
  - (4) little use of modern data analytics methods for decision making; and
  - (5) slow and cumbersome reporting features and lack of real-time monitoring.
- Data and analytics constraints:
  - PFM data often incomplete because of silos or lack of standardized storage and exchange.
  - Analysis and forecasting often rely on linear models, simple indicators, and traditional data sources, failing to leverage advanced analytics or AI.
  - Significant data go unused where agencies lack tools, licenses, capacity, or expertise to manage and interpret large databases.
- Resource and capacity gaps:
  - Less-developed economies face lack of financial resources, digital infrastructure, and required skill sets, widening the digital divide.
- Notable numeric reference:
  - The total cost of 156 FMIS projects financed by the World Bank since 1984, including government and borrower co-financing, has reached over $6 billion as of July 2024.

### B. Potential Future Scenarios of Digital PFM (futures cone)
- Scenario 1 — probable future (business as usual):
  - Legacy systems remain, slow adaptation, incremental improvements, responsive measures to maintain operations.
- Scenario 2 — plausible future (progressive modernization):
  - Upgrades to enhance cybersecurity, automate certain financial processes, expand interoperability, and incorporate advanced data analytics and AI assistance.
- Scenario 3 — possible future (transformative overhaul):
  - Comprehensive implementation of emerging technologies including AI for budget planning, tokenization for asset management, smart contracts, IoT, Web3 for digital identity, and digital money for payments and receipts.
  - Requires digital maturity, effective change management, and fundamental changes across institutions.
- Strategic considerations:
  - Radical overhaul has greatest potential impact but is most demanding; plausible future offers a balanced, pragmatic path.
  - Hybrid approach: adopt radical vision selectively for high-priority PFM functions or pain points.
- Timeline insight:
  - Radical overhaul of PFM systems will realistically span 10 to 20 years.
  - Early adopters or countries with strong digital foundations and leadership may achieve this future in less than a decade.

### C. Organizational approaches to adopting technology (overview)
- Value-Maximization (VM) approach: objective-driven technology adoption based on potential gain and feasibility; does not recommend specific technologies.
- Adoption decisions should consider:
  - specific PFM processes and tasks;
  - maturity and speculative enthusiasm surrounding each technology; and
  - unique context and culture of each government entity.

*International Monetary Fund. Harnessing Emerging Digital Technologies toward a New Frontier of Public Financial Management — Introduction (content unit: tnmea2026006 - Introduction2).*

### 1. Technological readiness: Emerging technologies such as AI, IoT, smart contracts, and digital money are rapidly advancing

- Key findings on readiness and timelines:
  - Widespread adoption in PFM requires maturity and proven use in PFM contexts.
  - Significant progress is expected within "5 to 10 years."
  - Some governments may take "more than 15 years" to develop comparable political will and leadership.
  - Building institutional capacity could take "5 to more than 20 years," depending on starting point, economy size, and investment in capacity building.
  - Fragile and conflict-affected states may find it difficult to achieve the vision within the "next 10 years."
  - Technology investments should be treated as ongoing strategic endeavors rather than one-off projects.

- Political leadership and policy:
  - Prioritization of digital transformation and commitment to innovation can drive rapid overhaul in PFM.
  - Global recognition of digital transformation benefits is growing and could accelerate progress.

- Resource and financing constraints:
  - Financing digital PFM initiatives is a significant challenge amid competing priorities.
  - Governments should secure funding through national budgets, international aid, or public–private partnerships.

- Organizational approaches and adopter types:
  - MoFs face a trade-off between early innovation (benefits sooner) and lowering disruption risks.
  - Four adopter types: passive adopters, reactive adopters, proactive adopters, and pioneers; a fifth category of nonadopters is noted.
  - Adopter types are fluid; organizations can evolve across types over time.

- Characteristics of adopter types:
  - Passive adopters: conservative, constrained by resources, skills, legacy systems; adopt mainly when driven by external pressures.
  - Reactive adopters: risk averse but open to incremental investments when benefits are demonstrated.
  - Proactive adopters and pioneers: actively explore digital trends; pioneers are experimental and resourceful.

- Illustrative country examples (Box 2):
  - Brazil:
    - National Treasury Secretariat modernized financial systems and payment processes using Pix.
    - Exploring Drex (Brazilian digital currency pilot) and ran a hackathon in "2023" to develop solutions using tokenized government securities.
  - Philippines:
    - Bureau of the Treasury (BTr) sold its first tokenized bonds in "November 2023," raising "$270 million" from institutional buyers.
    - Built on efforts like the Bonds.ph app and collaboration with state-owned banks and Philippine Digital Asset Exchange; used a dual registry structure to navigate compatibility issues.

- Trends and global indicators:
  - By "2025," "112" countries—more than half of the "198" surveyed countries—had introduced a national strategy on emerging technologies.
  - Governments increasingly adopt cloud, enterprise architecture, interoperability frameworks, and open-source policies.
  - As of "2022," less than a quarter of budget and accounting departments in OECD countries were using AI and RPA and cloud computing; a majority reported considering implementation.
  - By "2025," more than "100" countries have approved a national strategy on AI, and another "20" countries plan to approve over the next three years.

### 1. Central FMIS modernization / 2. Interest in technology developments — adoption evidence and examples

- Major findings on adoption and interest:
  - The rate of adoption for AI and RPA and cloud computing for ministries of finance in OECD countries was about 23.3 percent, respectively.
  - Comparative adoption figures cited:
    - Global adoption rate of government cloud platform at 58 percent in 2022.
    - Share of European enterprises buying cloud computing services at 41 percent in 2021.
    - Available estimates of AI and RPA adoption in the private sector for similar years range from 34 percent to more than 50 percent.

- Technologies and practices categorized by primary PFM purpose (capture/exchange/store; automation/control; architecture/cybersecurity; data exploitation; transparency):
  - Capture, exchange, and store:
    - Relational DBMS (MySQL, PostgreSQL); Distributed Ledger Technologies for automated settlement of public debt instruments.
    - Rwanda exploring machine learning for budget line classification; Brazil used natural language processing to reduce classification time from 1,000 hours to only 8 hours.
    - Cloud computing adopted by Brazil, Chile, and India for select FMIS modules; experimentation with data lakes.
    - API-first approach led by United Arab Emirates and United Kingdom.
  - Automation and control:
    - RPA in Finland to verify and reconcile expenditure; United States Treasury explored smart contracts for drawdown approvals.
    - RPA for electronic invoicing in Portugal; smart contracts explored in a Dutch municipality.
    - Kazakhstan piloting central bank digital currency for a subsidy program; digital identity technologies (Open Authorization, OpenID Connect) for beneficiary verification.
  - System architecture, deployment, cybersecurity:
    - Paraguay’s FMIS built on microservices (Java, Spring, Docker).
    - Honduras considering Kubernetes; uses Hashicorp Vault and CI/CD security integration via Jenkins and ArgoCD; DevSecOps and Zero Trust Security principles.
    - Zero-knowledge proof example: enable G2P payments without exposing recipient’s age, income, amount, and use of funds.
  - Data exploitation and decision making:
    - Business intelligence and data visualization tools; AI for augmented analytics and agentic AI.
    - Geographic Information System applications in India for welfare program monitoring and audit evidence.
    - AI and machine learning for macrofiscal and cash flow forecasting and nowcasting.
    - Georgia prototype AI application for Treasury payments where 42 machine learning models analyze historical data to detect anomalies in the “green channel”.
    - Generative AI used for producing and formatting documents; Brazil’s Treasury exploring generative capabilities for contracts and technical notes related to debt.
  - Fiscal transparency:
    - Fiscal transparency portals, mixed/augmented reality, data storytelling, and visualization.
    - Mexico publishes public investments portfolio in georeferenced platforms for monitoring public works.
    - South Africa and Uruguay publish budget and spending data using the Open Fiscal Data Package and APIs.

- Technology maturity, hype, and implications:
  - Maturity influences feasibility; nascent technologies may have flaws that reduce over time.
  - Technology S curve and Gartner’s Technology Hype Cycle guide timing decisions.
  - Example nuance: computer vision on Gartner’s plateau of productivity but limited PFM use cases.

### VM (Value Maximization) approach for decision making — formulas, scoring, and guidance
- Core formula:
  - VM = G × F
- Gain (G): score 1 to 10; weighted sum across six gain areas (Equation (2)):
  - G = ∑_{j=1} (G_j × W^G_j) = (G_P × W^G_P) + (G_OP × W^G_OP) + (G_OC × W^G_OC) + (G_A × W^G_A) + (G_S × W^G_S) + (G_F × W^G_F)
  - Six gain areas: process (G_P), output (G_OP), outcome (G_OC), accountability (G_A), security (G_S), futureproofing (G_F).
  - Weights W^G_j sum to 1.
- Feasibility (F): score 1 to 10; weighted sum across four feasibility factors (Equation (3)):
  - F = ∑_{k=1} (F_k × W^F_k) = (F_T × W^F_T) + (F_E × W^F_E) + (F_O × W^F_O) + (F_S × W^F_S)
  - Four feasibility factors: technical (F_T), economic (F_E), operational (F_O), scheduling (F_S).
  - Weights W^F_k sum to 1.
- Guidance on indicators:
  - Use SMART indicators and baselines (examples: percentage improvement in accuracy of fiscal forecasting; reduction in time overruns; speed of decision making; percentage reduction in time spent on routine tasks; reduction in cybersecurity incidents).
  - Feasibility considerations: compatibility, maturity, ROI, disruptions, training, legal changes, realistic timelines.

### Feasibility factors, risks, and mitigation strategies
- Feasibility factors: Technical, Economic, Operational, Scheduling.
- Key risks and mitigations:
  - Cybersecurity risks: mitigation via comprehensive cybersecurity frameworks (for example, National Institute of Standards and Technology, ISO 27001), regular vulnerability assessments, strong access controls and encryption.
  - Data privacy and sovereignty risks: mitigation via robust data governance, encryption, data privacy assessments, legal compliance review, data localization strategies.
  - Legal and regulatory risks: mitigation via policy frameworks and legal engagement upfront.
  - Skill and knowledge gaps: mitigation via continuous training, capacity building, and knowledge management systems.
  - Ethical risks: mitigation via code of ethics, ethics committees, ethics audits, and ethics training.
  - Change management risks: mitigation via comprehensive change management plans and early engagement.
  - Vendor lock-in risks: mitigation via multivendor strategies, data portability, open standards, flexible licensing and exit clauses.
  - Interoperability concerns: mitigation via data standards, interoperability testing, and vendor collaboration.

### Decision support tools, parameters, and scenarios
- Parameters for VM (component ranges and thresholds):
  - Gain (1–10): Low threshold ≤4 (example: cost savings of less than 5 percent); High threshold ≥7 (example: reduction of 50 percent in processing time).
  - Feasibility (1–10): Low threshold ≤4 (example: downtime of 20 percent expected); High threshold ≥7 (example: strong compatibility with existing systems).
  - Value (1–10): 0 < 30 indicates Low gain or feasibility; >50 indicates High gain or feasibility.
- Gain–Feasibility matrix quadrant guidance:
  - I. High gain and high feasibility: consider adoption.
  - II. High gain and low feasibility: consider adoption with caution (phased/pilot).
  - III. Low gain and low feasibility: wait for maturity or institutional changes.
  - IV. Low gain and high feasibility: adopt with managed expectations.
- Decision tree: stepwise assessment on whether reforms can solve the problem, presence of innovation labs/sandboxes, infrastructure and skills readiness, technology-agnostic theory of change, maturity evidence, measurable gains, and pilot/testing recommendations.

### Governance and sustainable adoption practices — 10 practical considerations
- Ten practices:
  1. Establish a clear vision and strategy.
  2. Adhere to interoperability and standards.
  3. Secure sustainable funding.
  4. Anticipate and manage resistance.
  5. Embrace agility and iterative development.
  6. Emphasize user experience through engagement.
  7. Build capacity and skills; retain them.
  8. Collaborate and seek partnerships.
  9. Mitigate security feasibility and ensure data privacy.
  10. Keep an eye out for new emerging technologies.
- Implementation notes:
  - Reengineer business processes and pursue legal reforms alongside technology adoption.
  - Use innovation labs and sandboxes to reduce early-adoption risks.
  - Post-implementation validation and feedback loops to compare realized outcomes against predicted scores and adjust future weighting.

*Source: Harnessing Emerging Digital Technologies toward a New Frontier of Public Financial Management — Technical Notes and Manuals (IMF).*

### Annex 1. Exploring New Technologies — labs, sandboxes, and testbeds
- Structured frameworks (innovation labs, regulatory sandboxes, technology testbeds) provide controlled, risk-mitigated environments for experimentation.
- Government innovation labs:
  - Serve as incubators drawing technologists, academics, and industry specialists.
  - Example: US 18F Lab assisted 34 agencies in completing 455 projects in 10 years to improve government digital services.
- Regulatory sandboxes:
  - Allow testing of innovative products/services under relaxed regulatory requirements (useful in fintech and related domains).
  - Examples: Dubai Financial Services Authority’s Innovation Testing License Program; Monetary Authority of Singapore’s fintech regulatory sandbox.
- Technology testbeds:
  - Dedicated environments to test specific technologies (blockchain, AI).
  - Example: Japan’s RIKEN Center for Advanced Intelligence Project for AI research and industry collaboration.
- Practical implications:
  - Use phased approaches: pilot in labs/testbeds, apply regulatory sandboxes for regulated domains, scale only after demonstrated interoperability, legal clarity, and cost–benefit outcomes.

### Annex 3. Practices for Sustainable Adoption of Emerging Technologies — leadership, people, and security
- Leadership and vision:
  - Adopt a forward-looking culture; incentivize personnel to learn and embrace new processes; recognize institutional constraints and legal hurdles.
- Emphasize user experience through engagement:
  - Involve end users in design, testing, and feedback loops; integrate monitoring and evaluation.
- Secure sustainable funding:
  - Establish dedicated funding mechanisms; seek private sector or international partnerships; advocate for budget allocations prioritizing technology modernization.
- Anticipate and manage resistance:
  - Deploy digitalization champions and change leaders to drive cultural change and ensure timely decisions.
- Build—and retain—capacity and skills:
  - Prioritize staff training and retention strategies; mitigate vendor lock-in and address legacy system integration strategically.
- Adhere to interoperability and standards:
  - Treat data governance, quality, and standardization as fundamental; leverage subject matter expertise for big data and AI systems.
- Embrace agility and iterative development:
  - Use agile and iterative methodologies adapted to public sector constraints; document and share progress openly.
- Collaborate and seek partnerships:
  - Work with other agencies, private sector, academia, and civil society; engage startups for agility and innovation.
- Mitigate security risks and ensure data privacy:
  - Invest in cybersecurity measures and a PFM cybersecurity risk mitigation strategy covering infrastructure vulnerability, social engineering, encryption stewardship, diffused responsibility, logging/audit trails, and periodic testing.
  - Recognize increased attack surfaces with system capability and complexity; address big data privacy risks and regulatory/ethical concerns.
- Keep an eye out for new technologies and cloud adoption:
  - Newer technologies reduce barriers to entry and can be implemented on cloud pay-as-you-go models; cloud deployment reduces need for data centers but requires robust internet connectivity and attention to connectivity limitations.

*International Monetary Fund. Annex 3. Practices for Sustainable Adoption of Emerging Technologies.*

### Introduction2

### Introduction2

### I. Envisioning Digital Transformation of PFM
- Vision: a digital PFM ecosystem built on cutting-edge digital solutions enabling real-time, data-driven decision making to achieve core PFM objectives:
  - (1) fiscal discipline—ensuring the overall control and sustainability of public finances;
  - (2) allocative efficiency—allocating resources according to government priorities;
  - (3) operational efficiency—executing government programs in an efficient manner to deliver public goods and services; and
  - (4) transparency and accountability—disclosing timely, comprehensive, and accurate fiscal information and institutionalizing accountability in the management of public funds.
- Potential benefits of properly harnessed emerging digital technologies:
  - increase the efficiency and effectiveness by automating routine tasks;
  - facilitate data-driven decision making across the budget cycle;
  - ensure fiscal sustainability by supporting better data analytics and decisions in fiscal policymaking;
  - strengthen transparency and accountability by allowing stakeholders to monitor the use of public funds;
  - improve public service delivery by enabling seamless, real-time interactions between citizens and government agencies; and
  - promote compliance with global norms through the adoption of international standards and practices in the PFM area.
- Examples of possible redesigns "digital by design":
  - Budget preparation: AI agents automating data validation, reporting, and workflow approvals, reducing manual effort and enabling faster, evidence-based budget formulation.
  - Public investment management: building information modeling tools; light detection and ranging technologies; and AI-driven analytics for planning, monitoring, and timely completion.
  - Government payments: new payment systems and digital money innovations to process payroll and subsidies rapidly and reach financially excluded populations.
  - Government contracts: smart contracts and IoT-enabled asset registries to streamline initiation, automate execution, and ensure transparent procurement oversight.
  - Public debt management: tokenization, distributed ledger technologies, and central bank digital currencies to enable issuance and automated settlement of digitalized debt instruments.
  - Performance budgeting: machine learning for target analysis and generative AI to produce tailored reports.
  - Internal control and audit: AI and analytics tools to detect anomalies, noncompliant and fraudulent transactions, and facilitate proactive, adaptive, real-time risk management.

### A. Current Digital PFM Landscape
- Coverage and systems:
  - Almost all countries use automated systems to support PFM functions, typically in the form of an FMIS supporting budget planning and preparation, budget execution, treasury operations, cash and debt management, public investment management, and accounting and reporting.
  - Additional functions—tax collection, debt management, payroll management, and procurement—are generally automated in separate, stand-alone systems.
- Key shortcomings of many existing PFM digital solutions:
  - (1) poor quality or incomplete, inconsistent data;
  - (2) silos or lack of interoperability across databases, systems, and institutions;
  - (3) outdated IT architecture with limited flexibility, scalability, and security;
  - (4) little use of modern data analytics methods for decision making; and
  - (5) slow and cumbersome reporting features and lack of real-time monitoring.
- Data and analytics constraints:
  - Cross-country surveys indicate data quality remains a concern; PFM data are often incomplete because of silos or lack of standardized storage and exchange.
  - Analysis and forecasting often rely on linear models, simple indicators, and only traditional data sources, failing to leverage advanced analytics or AI.
  - In some cases, significant data go unused because agencies lack tools, licenses, capacity, or expertise to manage and interpret large databases.
- Resource and capacity gaps:
  - Less-developed economies face additional challenges: lack of financial resources, digital infrastructure, and required skill set, widening the digital divide and making PFM transformation more difficult.
- Notable numeric references:
  - The total cost of 156 FMIS projects financed by the World Bank since 1984, including government and borrower co-financing, has reached over $6 billion as of July 2024.

### B. Potential Future Scenarios of Digital PFM
- Framework: futures cone mapping scenarios by likelihood—probable, plausible, and possible.
- Scenario 1 — probable future:
  - Business as usual: legacy systems remain, slow adaptation, incremental improvements, and responsive measures to maintain operations.
- Scenario 2 — plausible future:
  - Progressive modernization recognizing current trends: upgrades to enhance cybersecurity, automate certain financial processes, expand interoperability, and incorporate advanced data analytics and AI assistance.
- Scenario 3 — possible future (transformative overhaul):
  - Comprehensive implementation of emerging technologies, including:
    - AI for budget planning and reporting;
    - tokenization for asset management;
    - smart contracts for payroll, debt, and procurement management;
    - IoT for inventory and public investment management;
    - Web3 for digital identity; and
    - digital money for payments and receipts.
  - Requires significant and fundamental changes, digital maturity, and effective change management.
- Strategic considerations:
  - The radical overhaul may have the greatest potential impact but is demanding; the plausible future offers a balanced, pragmatic path with substantial advantages.
  - A hybrid approach—adopting a radical vision for select PFM functions or processes according to priorities and pain points—can drive meaningful progress.
- Timeline insight (from Box 1):
  - The timeline for a radical overhaul of PFM systems, partial or full, depends on several factors but will realistically span 10 to 20 years.
  - Early adopters or countries with strong digital foundations and leadership may achieve this future in less than a decade.

### C. Organizational Approaches to Adopting Technology (overview)
- Strategic adoption should reflect benefits, costs, risks, and limitations.
- The note proposes a value-maximization (VM) approach: objective-driven technology adoption based on potential gain and feasibility, without recommending specific technologies.
- Adoption decisions should consider:
  - specific PFM processes and tasks;
  - the maturity and speculative enthusiasm surrounding each technology; and
  - the unique context and culture of each government entity.

*International Monetary Fund. Harnessing Emerging Digital Technologies toward a New Frontier of Public Financial Management — Introduction (content unit: tnmea2026006 - Introduction2).*

### 1. Technological  readiness:  Emerging  technologies  such  as  AI,  IoT,  smart  contracts,  and  digital

### 1. Technological readiness: Emerging technologies such as AI, IoT, smart contracts, and digital money are rapidly advancing

### Key findings on readiness and timelines
- Emerging technologies are rapidly advancing but widespread adoption in PFM requires maturity and proven use in PFM contexts.
- Significant progress is expected within "5 to 10 years."
- Some governments may take "more than 15 years" to develop comparable levels of political will and leadership.
- Building institutional capacity could take "5 to more than 20 years," depending on starting point, economy size, and investment in capacity building.
- Fragile and conflict-affected states may find it difficult to achieve the vision within the "next 10 years."
- Technology investments should be treated as ongoing strategic endeavors rather than one-off projects.

### Political leadership and policy
- Prioritization of digital transformation in policy agendas and a commitment to innovation can drive rapid overhaul in PFM.
- Global recognition of the benefits of digital transformation is growing and could accelerate progress.

### Resource and financing constraints
- Financing digital PFM initiatives is a significant challenge amid competing priorities.
- Governments should secure funding through national budgets, international aid, or public–private partnerships.

### Organizational approaches to adopting technology
- MoFs face a trade-off between early innovation (benefits sooner) and lowering disruption risks.
- Four adopter types are distinguished: passive adopters, reactive adopters, proactive adopters, and pioneers; a fifth category of nonadopters is noted but not discussed in depth.
- Adopter types are neither mutually exclusive nor fixed; organizations can evolve across types over time.

### Characteristics of adopter types
- Passive adopters (aligned with business-as-usual future):
  - Conservative culture prioritizing stability, continuity, reliability.
  - Likely constraints: lack of financial resources, lack of skilled personnel, complex legacy systems, resistance to change.
  - Adopt digital modernization mainly when driven by external pressures (laws/regulations).
- Reactive adopters (aligned with progressive modernization future):
  - Risk averse but open to innovation.
  - Prefer case-by-case, incremental investments when risks are low or benefits are demonstrated.
- Proactive adopters and pioneers (aligned with a transformative overhaul future):
  - Open to innovation and actively explore digital trends and good practices.
  - Frequently adopt technologies tested elsewhere.
  - Pioneers are most experimental, willing to take risks, and possess resources for experimentation (well-developed digital infrastructure, skilled workforce, supportive leadership).

### Cultural and institutional dynamics
- Organizational culture, resource availability, and perceived risks influence innovation approach and likely future scenario realization.
- PFM structures, institutions, and processes should be flexible and adaptable to enable continuous automation and digitalization.
- The digital journey in PFM is framed as an ongoing evolution, not a one-time reform.

### Illustrative country examples (Box 2)
- Brazil:
  - National Treasury Secretariat has modernized financial systems and payment processes using Pix.
  - Exploring Drex (Brazilian digital currency pilot) and ran a hackathon in "2023" to develop solutions using tokenized government securities.
  - Emphasis on an organizational culture of agility, collaboration, innovation, and data use.
- Philippines:
  - Bureau of the Treasury (BTr) sold its first tokenized bonds in "November 2023," raising "$270 million" from institutional buyers.
  - Built on efforts like the Bonds.ph app and collaboration with state-owned banks and Philippine Digital Asset Exchange.
  - Used a dual registry structure for securities to navigate compatibility issues via private-sector collaboration.
  - Aiming to leverage widespread use of the GCash app to offer tokenized bonds to small retail investors and promote financial inclusion.

### Trends in adoption and global indicators
- By "2025," "112" countries—more than half of the "198" surveyed countries—had introduced a national strategy on emerging technologies.
- Governments are increasingly adopting cloud, enterprise architecture, interoperability frameworks, and open-source policies.
- As of "2022," less than a quarter of budget and accounting departments in OECD countries were using AI and RPA and cloud computing; a majority reported considering implementation.
- Anecdotal evidence: MoFs explore AI mainly as an assistive tool to improve existing processes.
- The 2022 OECD survey predates the recent rise of generative and agentic AI.
- As of "2025," more than "100" countries have approved a national strategy on AI, and another "20" countries plan to approve over the next three years.

*Source: Authors.*

### 1. Central FMIS modernization2. Interest in technology developments

### 1. Central FMIS modernization2. Interest in technology developments

### Major findings on adoption and interest
- The rate of adoption for artificial intelligence and robotic process automation (AI and RPA) and cloud computing for ministries of finance in Organisation for Economic Co-operation and Development countries was about 23.3 percent, respectively.
- Comparative reference adoption figures (not directly comparable) cited:
  - Global adoption rate of government cloud platform at 58 percent in 2022 (World Bank 2025).
  - Share of European enterprises buying cloud computing services at 41 percent in 2021 (Eurostat 2023).
  - Available estimates of AI and RPA adoption in the private sector for similar years range from 34 percent to more than 50 percent.
- Specific country data gaps noted: Chile, Colombia, Israel, Mexico, Slovenia, and the United States data are not available. For cloud-based computing services, Czechia and Iceland data are not available.

### Technologies and practices categorized by primary PFM purpose (capture, exchange, store; automation and control; architecture/cybersecurity; data exploitation; transparency)
- Capture, exchange, and store data (Table 1) — examples and findings:
  - Recording data / Accounting and financial reporting: relational database management systems such as MySQL and PostgreSQL; Distributed Ledger Technologies can be integrated for automated settlement of public debt instruments.
  - Classifying data / Fiscal forecasting and budget preparation: content classification systems and AI; examples: Rwanda exploring machine learning for budget line classification; Brazil used natural language processing to reduce classification time from 1,000 hours to only 8 hours.
  - Storing, accessing, retrieving data: cloud computing enables scalable storage and access; Brazil, Chile, and India adopted cloud-based solutions for select FMIS modules; experimentation with data lakes.
  - Exchanging data: API-first approach and data standards (national and international); examples: United Arab Emirates and United Kingdom spearheaded API-first initiatives.

- Automation and control (Table 2) — examples:
  - Controlling against a rule: RPA used in Finland to verify and reconcile expenditure; United States Treasury explored smart contracts for drawdown approvals and limits; AI for detecting potentially noncompliant spending.
  - Authorizing/validating: RPA for electronic invoicing in Portugal; smart contracts explored in a Dutch municipality for eligibility enforcement; computer vision/optical character recognition for scanned documents.
  - Payments and transfers: Kazakhstan piloting central bank digital currency for a subsidy program; digital identity technologies (Open Authorization, OpenID Connect) for beneficiary identification and verification.

- System architecture, deployment, cybersecurity (Table 3) — examples:
  - Building supportive digital ecosystem and IT architecture: Paraguay’s FMIS built on microservices; business layer developed using Java and Spring and deployed using Docker.
  - Deploying digital solutions: Honduras considering Kubernetes; use of orchestrators for microservices within cloud environments.
  - Securing PFM systems: Honduras uses Hashicorp Vault and integrates security checks into CI/CD via Jenkins and ArgoCD, aligning with DevSecOps and Zero Trust Security principles.
  - Ensuring data privacy: Zero-knowledge proof can enable validation without exposing underlying data (example: enabling G2P payments without exposing recipient’s age, income, amount, and use of funds).

- Data exploitation and decision making (Table 4) — examples:
  - Analyzing data for insights: business intelligence and data visualization tools used by ministries of finance; AI enables augmented analytics and agentic AI for orchestrating analyses; Geographic Information System applications in India for welfare program monitoring and audit evidence.
  - Projecting/forecasting: AI and machine learning to enhance macrofiscal and cash flow forecasting and to nowcast using unstructured, high-frequency data.
  - Assessing risks and scenarios: Georgia developed a prototype AI application for Treasury payments where 42 machine learning models analyze historical data to detect anomalies in the “green channel”.
  - Creating and updating content: generative AI used for producing and formatting documents; Brazil’s Treasury exploring these capabilities (contracts and technical notes related to debt under the Salary Variations Compensation).

- Fiscal transparency (Table 5) — examples:
  - Making information accessible: fiscal transparency portals with mixed/augmented reality, data storytelling, and visualization; Mexico publishes public investments portfolio in georeferenced platforms for monitoring public works using Geographic Information Systems.
  - Publication of open data: South Africa and Uruguay publish budget and spending data using the Open Fiscal Data Package; use of Application Programming Interface to ensure open access.

### Technology maturity, hype, and implications
- Maturity influences feasibility and adoption; nascent technologies may carry inherent flaws that reduce over time as they mature.
- Technology S curve concept: performance initially low then grows rapidly as adoption deepens; governments must time transitions appropriately.
- Gartner’s Technology Hype Cycle: innovation trigger → peak of inflated expectations → trough of disillusionment → slope of enlightenment → plateau of productivity.
- Example nuance: computer vision is on Gartner’s plateau of productivity but has limited PFM use cases; hence industry-level maturity assessments can miss PFM-specific applicability.

### VM (Value Maximization) approach for decision making (formulas and components)
- Core formula: VM = G × F
  - G: score for potential gains of technology adoption, range 1 to 10; weighted sum across gain areas (Equation (2)):
    - G = ∑_{j=1} (G_j × W^G_j) = (G_P × W^G_P) + (G_OP × W^G_OP) + (G_OC × W^G_OC) + (G_A × W^G_A) + (G_S × W^G_S) + (G_F × W^G_F)
    - Six gain areas: process (G_P), output (G_OP), outcome (G_OC), accountability (G_A), security (G_S), futureproofing (G_F).
    - Weights W^G_j sum to 1.
  - F: score for practical feasibility of adoption, range 1 to 10; weighted sum across feasibility factors (Equation (3)):
    - F = ∑_{k=1} (F_k × W^F_k) = (F_T × W^F_T) + (F_E × W^F_E) + (F_O × W^F_O) + (F_S × W^F_S)
    - Four feasibility factors: technical (F_T), economic (F_E), operational (F_O), scheduling (F_S).
    - Weights W^F_k sum to 1.

- Guidance on scoring and indicators:
  - Use SMART indicators and baselines to measure gains (examples in Table 6: percentage improvement in accuracy of fiscal forecasting, reduction in time overruns, speed of decision making, percentage reduction in time spent on routine tasks, reduction in cybersecurity incidents).
  - Feasibility should consider compatibility, maturity, ROI, disruptions, training, legal changes, and realistic timelines.

### Feasibility factors, risks, and mitigation strategies (Table 7 highlights)
- Feasibility factors: Technical, Economic, Operational, Scheduling.
- Key risks and example mitigations:
  - Cybersecurity risks (technical): data breaches, malware, phishing; mitigation: comprehensive cybersecurity framework (for example, National Institute of Standards and Technology, ISO 27001), regular vulnerability assessments, strong access controls and encryption.
  - Data privacy and sovereignty risks (technical): misuse of personal data, concerns over foreign storage, potential surveillance; mitigation: robust data governance, encryption, data privacy assessments, legal compliance review, data localization strategies if necessary.
  - Legal and regulatory risks (operational): lack of legal frameworks, compliance difficulties; mitigation: establish policy frameworks, engage legal experts upfront, stay updated with laws and regulations.
  - Skill and knowledge gaps (operational): shortage of qualified personnel; mitigation: continuous training and capacity building, collaborate with stakeholders, develop knowledge management systems.
  - Ethical risks (operational): algorithmic bias, lack of transparency; mitigation: develop a code of ethics, establish ethics committees or advisory bodies, conduct and publish ethics audits, provide ethics training.
  - Change management risks (operational): resistance to change; mitigation: comprehensive change management plan, engage employees early, provide communications and support.
  - Vendor lock-in risks (economic, operational): dependence on a single provider; mitigation: multivendor strategy, ensure data portability and interoperability through open standards, negotiate flexible licensing and exit clauses.
  - Interoperability concerns (technical, operational): integration with legacy systems; mitigation: adopt data standards, conduct interoperability testing, collaborate with vendors for compatibility.

### Decision support tools, parameters, and scenarios
- Parameters for Value Maximization (Table 8) — component ranges and thresholds:
  - Gain (1–10): Low threshold ≤4 (example: cost savings of less than 5 percent; no improvement expected in transparency scores); High threshold ≥7 (example: reduction of 50 percent in processing time; protection against data breach and business continuity).
  - Feasibility (1–10): Low threshold ≤4 (example: downtime of 20 percent expected; more than one year to implement); High threshold ≥7 (example: strong compatibility with existing systems; clear and positive return on investment).
  - Value (1–10): 0 < 30 indicates Low gain or feasibility; >50 indicates High gain or feasibility.
- Gain–Feasibility matrix (Figure 9) — quadrant guidance:
  - I. High gain and high feasibility: consider adoption.
  - II. High gain and low feasibility: consider adoption with caution (phased/pilot approach).
  - III. Low gain and low feasibility: consider waiting for technology maturity or institutional changes.
  - IV. Low gain and high feasibility: consider adoption with managed expectations.
- Decision tree (Figure 10): stepwise questions on whether reforms can solve the problem, availability of innovation labs/sandboxes, infrastructure and skills readiness, theory of change being technology-agnostic, maturity evidence, measurable gains, and pilot/testing recommendations.

### Governance and sustainable adoption practices (10 practical considerations)
- Ten practices for sustainable adoption (Figure 11):
  1. Establish a clear vision and strategy.
  2. Adhere to interoperability and standards.
  3. Secure sustainable funding.
  4. Anticipate and manage resistance.
  5. Embrace agility and iterative development.
  6. Emphasize user experience through engagement.
  7. Build capacity and skills; retain them.
  8. Collaborate and seek partnerships.
  9. Mitigate security feasibility and ensure data privacy.
  10. Keep an eye out for new emerging technologies.
- Additional implementation notes:
  - Reengineering business processes and legal reforms should accompany technology adoption to avoid automating inefficient manual tasks.
  - Innovation labs and sandbox approaches can reduce risks associated with early adoption.
  - Post-implementation validation and feedback loops are recommended to compare realized outcomes against predicted scores and adjust weights and scores for future projects.

*Source: Harnessing Emerging Digital Technologies toward a New Frontier of Public Financial Management — Technical Notes and Manuals (IMF).*

### Annex 1. Exploring New Technologies

### Annex 1. Exploring New Technologies

### Overview
- Structured frameworks such as innovation labs, regulatory sandboxes, and technology testbeds provide a controlled, risk-mitigated environment for experimentation and assessment of emerging technologies.
- These platforms facilitate a symbiotic relationship between and across government agencies and the private sector.

### Government Innovation Labs: A Multidisciplinary Approach
- Purpose: Serve as incubators for technological exploration and solution development, drawing upon the expertise of technologists, academics, and industry specialists; function as collaborative spaces where innovative ideas can be developed, tested, and refined to address specific public sector challenges.
- Example: The US 18F Lab was established by the General Services Administration to help agencies develop user-centric digital services that improve the public’s interaction with the government.
  - Composition: designers, software engineers, strategists, and product managers.
  - Performance: In 10 years, they have assisted 34 agencies in completing 455 projects that improved government digital services.

### Regulatory Sandboxes: Facilitating Innovation within Regulated Sectors
- Purpose: Enable exploration of innovative products, services, or business models under relaxed regulatory requirements, under specific conditions; particularly popular in heavily regulated domains such as financial sectors.
- Role for public financial management: Public financial management agencies may consider such an environment when exploring solutions with private sector actors.
- Examples:
  - Dubai Financial Services Authority opened an Innovation Testing License Program to foster fintech innovations within controlled parameters, supporting the creation of new financial services and products.
  - The Monetary Authority of Singapore opened a fintech regulatory sandbox to provide a pragmatic framework for testing financial technologies, such as blockchain for cross-border payments, under regulatory oversight, fostering innovation while safeguarding systemic integrity.

### Technology Testbeds: Testing Specific Technologies
- Purpose: Governments or agencies can establish dedicated testbeds for specific technologies (for instance, to explore the use of blockchain, or artificial intelligence (AI)), providing a controlled environment for companies and researchers to experiment with new applications and use cases and allowing ministries to assess potential benefits and risks.
- Example: The government of Japan has established RIKEN Center for Advanced Intelligence Project to conduct research on AI and machine learning, and to collaborate with industry partners to develop AI applications in various domains.

### Practical implications and recommended use
- Use structured frameworks to:
  - Mitigate risks while experimenting with emerging technologies.
  - Enable collaboration between government, private sector, and academia.
  - Assess interoperability, governance, accountability, and feasibility before wide deployment.
- Consider phased approaches:
  - Pilot in innovation labs or testbeds.
  - Use regulatory sandboxes where regulatory flexibility is required for fintech or similar domains.
  - Scale only after demonstrated interoperability, legal/regulatory clarity, and clear cost–benefit outcomes.

*Annex 1. Exploring New Technologies — International Monetary Fund*

### Annex 3. Practices for Sustainable

### Annex 3. Practices for Sustainable Adoption of Emerging Technologies

### Leadership and vision
- Adopt a forward-looking culture to succeed in implementing new technologies.
- Be open to change; discard outdated processes and systems that are incapable of adapting.
- Incentivize personnel to learn and embrace new ways of doing things.
- Recognize difficulty in government settings: preference for established ways, institutional constraints (many codified into law), and the nontrivial task of navigating and modifying those constraints.

### Emphasize user experience through engagement
- Prioritize user experience at every stage of technology implementation in public financial management (PFM).
- Actively involve end users in design, testing, and feedback loops.
- Note: knowledge of the business process from the perspective of ministry of finance cannot substitute direct engagement with the users.
- Integrate monitoring and evaluation mechanisms to assess and iteratively refine user-engagement strategies.
- Early and continuous engagement tailors solutions, increases acceptance, and drives long-term success.

### Secure sustainable funding
- Recognize budget constraints as a significant hurdle for acquiring, implementing, and maintaining technologies.
- Secure sustainable funding sources by:
  - Establishing dedicated funding mechanisms for technology adoption.
  - Seeking partnerships with the private sector or international organizations.
  - Advocating for budget allocations that prioritize technology modernization in PFM processes.

### Anticipate and manage resistance
- Expect skepticism and resistance among government agencies and PFM stakeholders.
- Emphasize that digital transformation requires focus on people as well as technology.
- Deploy digitalization champions or change leaders to drive cultural change, advocate for technologies that enhance efficiency, agility, productivity, and innovation in PFM.
- Ensure change leaders help ensure the right decisions are taken in the right way, at the right time, for the right outcome to keep the PFM digitalization agenda on track.

### Build—and retain—capacity and skills
- Managers must be aware of the current technology landscape and how new capabilities can be leveraged.
- IT management can disseminate information, but functional staff must understand technology as applied to their roles.
- Prioritize staff training to ensure safe and efficient implementation of new technologies.
- Develop retention strategies for qualified staff in government settings, coordinated with human resources.
- Address the common issue where vendors install systems and depart, leaving IT staff without specific training; this can lead to data loss, erroneous data, compromised security, fraud, and theft.
- Mitigate vendor lock-in and dependency risks, including data ownership concerns.
- Tackle legacy system resistance to integration with modern technologies through a strategic, multidimensional approach addressing technical and organizational aspects.

### Adhere to interoperability and standards
- Treat data governance, quality, and standardization as fundamental; inconsistencies can compromise financial information accuracy and reliability.
- Address interoperability issues when integrating new systems with existing PFM platforms to enable comprehensive data analysis and decision making.
- Leverage subject matter expertise to make sense of data for operating big data or artificial intelligence systems.

### Embrace agility and iterative development
- Use flexible and adaptive approaches, such as agile and iterative development methodologies.
- Emphasize incremental progress, frequent feedback loops, and collaboration among cross-functional teams.
- Tailor agile practices to public sector constraints: bureaucratic processes, rigid procurement policies, and resistance to change.
- Modify traditional project management approaches to allow more flexibility and iterative development.
- Increase transparency by documenting and sharing progress, issues, and decisions openly.

### Collaborate and seek partnerships
- Collaborate with other government agencies, private sector partners, academia, and civil society to leverage complementary expertise, resources, and networks.
- Share resources such as data, infrastructure, and expertise to enhance the effect and efficiency of technology adoption initiatives.
- Pursue joint research and development, co-creation of solutions, and shared learning opportunities.
- Engage start-ups focused on government technology for agility, innovative solutions, and fresh perspectives to overcome traditional barriers and improve efficiency and transparency in financial operations.

### Mitigate security risks and ensure data privacy
- Invest in robust cybersecurity measures to protect sensitive financial information from unauthorized access and cyberattacks.
- Develop a cybersecurity risk mitigation strategy in the PFM area that accounts for:
  - Vulnerability of infrastructure to penetration and compromise
  - Vulnerability of staff to “social engineering”-type attacks
  - Proper stewardship and management of encryption and encrypted resources, especially encryption keys
  - Diffused responsibility for infrastructure management so that no one individual has complete control
  - Logging, audit trails, and attribution; actions should be attributable to an individual
  - Periodic testing of security mechanisms and controls against the latest exploits
- Recognize that true security demands expert attention; where budgets constrain hiring qualified security staff, consider funding a dedicated cybersecurity team at the government or ministry level to set and enforce security policy, conduct audits and testing, and educate staff in security awareness.
- Acknowledge that increased system capability and complexity enlarge the “attack surface,” increasing risks from implementor carelessness, off-the-shelf software bugs (for example, “zero-day exploits”), misconfiguration, or social engineering.
- Recognize big data systems can threaten individual privacy by enabling aggregation and inference across disparate data sets; integration may violate privacy laws and create risks of misuse and data leaks.
- Navigate intricacies of data privacy alongside regulatory frameworks, ethical concerns, public trust, and political and bureaucratic barriers.

### Keep an eye out for new emerging technologies and cloud adoption
- Note that newer technologies are characterized by lower barriers to entry and can often be implemented in the cloud on pay-as-you-go or pay-per-use models, reducing the need for physical processing infrastructure.
- Cloud deployment reduces the need to build data centers and purchase servers and ancillary equipment, and reduces the amount of support personnel required to maintain traditional data centers.
- Recognize that cloud-based technologies require robust internet connectivity for effective use; end users will need a certain level of internet connectivity.
- Understand that private networks used in traditional deployments are of limited use for cloud deployments; instead, a robust connection to the cloud provider (internet) is necessary.
- Expect elimination of specialized hardware at end user sites and expanded access methods, reducing the technical burden on network administrators at both end user and data center levels.

*International Monetary Fund. Annex 3. Practices for Sustainable Adoption of Emerging Technologies.*

---


_Source: https://www.imf.org/-/media/files/publications/tnm/2026/english/tnmea2026006.pdf_