Governments today face a growing disconnect between rising expectations for speed, adaptability and responsiveness, and institutional systems that have not kept pace. Digital technologies and data are no longer optional enablers; they have become core infrastructure for governments seeking to address today’s policy and service delivery challenges. Yet the 2025 OECD Digital Government Index and OURdata Index show that while progress is real, it remains concentrated in strategies, frameworks and enabling conditions, rather than in their consistent operational application. Governments have invested in many of the right foundations, but these often fail to function as an integrated system. Weak data governance limits coherence and reuse; digital public infrastructure is deployed but underused; investment, procurement and workforce systems remain too rigid for iterative digital delivery; artificial intelligence (AI) is advancing faster than the trust mechanisms needed to steward it; and services remain too reactive and fragmented to reach people in a simple and proactive way. This chapter sets out the evidence behind this diagnosis, identifies the main challenges preventing governments from moving from policy design to operational delivery, and proposes the actions needed to close the gap between digital ambition and strong public sector performance.
Digital Government Outlook 2026
From Foundations to Transformational Impact
Abstract
Efficient and adaptive governments remain essential. Around the world, governments operate in a context of rapid change and disruption, driving increasingly dynamic and uncertain policy environments. Governments simultaneously navigate geopolitical tension, economic volatility, demographic pressure, environmental challenges and rapid technological change. In such an environment, resilience is not only about responding to crisis. It requires governments to anticipate and to adjust quickly, learn continuously and reconfigure their institutions’ methods and processes at pace to protect people from evolving risks and ensure that services remain relevant to their needs.
People expect government to deliver solutions to increasingly complex and fast-moving challenges. These expectations are shaped by many factors, including people’s direct experience of public services and by their perceptions of how government responds to high-stakes, cross-cutting policy issues that affect long-term well-being. The 2024 OECD Survey on the Drivers of Trust in Public Institutions shows that people have limited confidence in governments’ capacity to address complex policy issues involving multiple trade-offs: while 77% of people believe adapting to automation and new technologies should be a high national priority, only 41% trust their government to regulate them appropriately. This is further compounded by a frenetic and fragmented information eco-system. Together, these dynamics make it more difficult for governments to communicate clearly and mobilise collective action in times of uncertainty.
Against this backdrop, the effective, coherent and trustworthy use of digital technologies and data is foundational for an efficient and adaptive government. Digital transformation enables governments to anticipate risks earlier, co-ordinate complex interventions and respond at pace. But technology alone is not enough. It also requires strong digital governance, robust data stewardship, a skilled and adaptable workforce and legal and ethical frameworks that safeguard trust. These elements form the core infrastructure of a public sector capable of better joining up government institutions, navigating uncertainty, delivering high‑quality services and strengthening confidence in the digital age.
However, many governments face structural constraints that limit their agility to respond to this changing environment and expectations. Tight budgets, obsolete rules and regulations, rigid processes, and fragmented institutional responsibilities slow decision-making and weaken co-ordination. While digital government strategies have grown more ambitious in recent years, governance and delivery mechanisms have not always kept pace. As a result, governments increasingly find themselves caught between rising expectations for speed, adaptability and responsiveness, and institutional systems not designed for rapid, co-ordinated action. This gap is the central challenge for governments today.
Closing the gap between rising expectations and institutional capacity requires more than technology adoption. It demands a fundamental re-engineering of how governments organise, govern and invest in their digital capabilities. Layered processes, rigid rules and fragmented systems have made many governments structurally slow. Connecting data, modernising systems and adopting more agile ways of working can make governments faster, more efficient and more responsive without sacrificing quality and accountability. The next phase of digital government reform requires governments to change how they organise, govern and invest in digital capability, with greater focus on embedding delivery capacity across institutions rather than expanding policy frameworks
The OECD Digital Government Policy Framework (DGPF) provides a structured, evidence-based approach to this challenge (OECD, 2020[1]). Building on the OECD Recommendation on Digital Government Strategies, the DGPF identifies six dimensions of digital maturity that governments must develop to deliver a coherent, human-centred and whole-of-government transformation (Figure 1.1). They are: (1) Digital by design; (2) Data-driven public sector; (3) Government as a platform; (4) Open by default; (5) User-driven; and (6) Proactiveness.
Source: (OECD, 2020[1]).
Digital by design means embedding digital thinking into the design of policies, processes and services from the outset, rather than adding technology onto legacy structures. It acts as a strategic lever across the public sector, creating the conditions for the other dimensions to take root. A government that is digital by design reconfigures how it operates at the source, rather than managing workarounds.
A data-driven public sector puts data at the centre of decision-making, service design and public accountability. Governments that systematically govern, share and use data can anticipate needs, reduce duplication and tailor services more precisely. This dimension is foundational to both efficiency and the capacity to absorb and respond to shocks.
Government as a platform shifts the model from siloed delivery to shared infrastructure: common building blocks, interoperability standards and reusable components that allow agencies to co-ordinate without rebuilding from scratch. This directly addresses the fragmentation that widens the gap between institutional capacity and what people and businesses expect.
Open by default moves governments away from closed, top-down decision-making towards transparency, data openness and collaborative engagement with people, civil society and the private sector. It enables governments to draw on broader knowledge and resources, and to build the trust that underpins legitimacy in the digital age.
A user-driven government centres the design and delivery of policies and services around the actual needs of people, addressing the digital divide and ensuring that digital transformation works for all. It requires systematic engagement with users across policy and service lifecycles, moving beyond formal consultation to continuous feedback.
Proactiveness goes further. It captures governments' capacity to anticipate users' needs and deliver services before they are explicitly requested, using the once-only principle and data analytics, AI and predictive tools to shift from reactive to anticipatory government.
To understand how countries are putting these six dimensions into practice, the Digital Government Policy Framework examines each one through four transversal facets that mirror the stages of the policy cycle: strategic approach, policy levers, implementation and monitoring. The policy cycle lens therefore offers a more targeted and actionable understanding of where reforms stand and, crucially, where intervention is most needed to turn digital maturity into tangible improvements in government performance.
Developing maturity across all six dimensions involves navigating trade-offs. Efficiency pushes governments towards consolidation, standardisation and leaner processes; resilience requires redundancy, fallback channels and the capacity to absorb shocks. Investment models and workforce strategies must also keep pace: flexible, staged funding allows governments to test, learn and scale iteratively, while multidisciplinary teams and continuous professional development build the capabilities that technology alone cannot substitute.
Many of the challenges governments face, such as pandemics, environmental disasters and geopolitical shocks, are increasingly cross‑border and cross‑sector in nature, compounding the gap that no single government can close alone. Shared digital standards, interoperable data infrastructures and cross-border digital credentials reduce duplication and enable whole-of-government responses under strain. Instruments such as eIDAS 2.0 and the European Digital Wallet support trusted access within and across borders. Partnerships with the private sector, academia and civil society further expand capacity for collaborative responses to transnational risks.
Measuring digital government maturity is thus essential to understand how countries can strengthen their digital capabilities and what may be impeding them to turn developments into greater government performance. Furthermore, benchmarking digital capabilities helps policymakers see where the building blocks of digital government are in place, identify strengths to build on, and pinpoint where weak implementation is leaving governments stuck and in need of greater investment or political attention. To support this, the OECD has developed two complementary indexes that together provide a comprehensive picture of digital government performance: the Digital Government Index (DGI) and the Open, Useful, and Re-usable Data Index (OURdata). Together, these indexes benchmark the enabling foundations for a coherent and human-centred government digital transformation across OECD Member and accession countries (Box 1.1).
Despite progress across countries, implementation remains uneven. Governments with frameworks but without sustained capability, shared platforms or assurance mechanisms struggle to translate intent into delivery at the pace and scale needed. Progress depends on translating digital ambition into institutional transformation, supported by operating models, incentives and accountability arrangements that enable sustained adoption and use. The remainder of this chapter examines where countries stand across each of the six dimensions, and what it will take to move from digital ambition to institutional transformation.
The Digital Government Index (DGI) assesses the enabling foundations for digital transformation across the six dimensions of the Digital Government Policy Framework (DGPF). It measures government actions on a wide range of critical enablers to strategically leverage digital technologies and data in government: governance and investments management, digital infrastructure, digital skills, data governance and sharing, artificial intelligence in government, and service design and delivery. It collects 155 datapoints per each participant country.
Progress across each dimension is assessed through four transversal facets that reflect successive stages of the policy cycle: strategic approach (the extent to which governments have coherent vision and direction); policy levers (the extent to which right mandates, legal frameworks and tools are in place); implementation (the extent to which these are embedded in operational practice); and monitoring (the extent to which governments evaluate results and learn). This four-stage lens is central to understanding not just how far governments have come, but where in the reform process progress has stalled.
The Open, Useful and Re-usable Data (OURdata) Index provides a specific benchmark on the robustness of open government data policies, examining the extent to which open government data (including high-value datasets) are available, accessible, as well as the degree governments proactively support their reuse. It collects 670 datapoints per each participant country. Selected datapoints are included in the calculation of the DGI, dimensions Data-driven public sector and Open by default.
What makes these two indexes distinctive is that they go beyond checking whether comprehensive digital government policies exist on paper, or government services are offered by digital means. They assess whether governments have the capability to implement those policies coherently across the public sector, from initial design through to final delivery. Therefore, they are essential instruments for OECD Member and partner countries to design their digital government strategies, benchmark progress, prioritise efforts and, overall, build more solid digital government capabilities.
Methodological process
Both indexes are the product of collaborative work with OECD Member countries through the Working Party of Senior Digital Government Officials (E-Leaders) to establish rigorous, transparent, and coherent measures, with their methodology approved by the Working Party and declassified by the OECD Public Governance Committee (PGC).
The data presented in this report draws on the OECD Survey on Digital Government 3.0, which informs the DGI; and on the OECD Survey on Open Government Data 6.0, which informs the OURdata Index. Both surveys were collected in the first half of 2025, covering policies and initiatives in place between 1 January 2023 and 31 December 2024.
Survey respondents were high-level digital government officials designated by each participating country, and a glossary of terms was provided to ensure consistent interpretation across jurisdictions. Once the data collection period closed, all country responses underwent a detailed, multi-round validation process. A first round reviewed responses for internal consistency and verified that answers and supporting evidence corresponded to the respective question. A second round ensured transversal consistency across survey sections and themes, identifying contradictions or gaps between related questions. Throughout this process, the OECD engaged bilaterally with country respondents to clarify ambiguities, request additional evidence and resolve discrepancies.
Following validation, each participating country formally reviewed and approved the final dataset before it was used to calculate the composite indices and inform the analysis presented in this report. This approval step ensures that the data accurately reflects each country's situation during the reference period and establishes a shared commitment to the integrity of the dataset.
Further methodological details, including the weighting scheme, aggregation method and statistical robustness tests, are presented in OECD working paper with composite results (2026[2]).
Source: OECD (2026[2]).
This report is based on the results and insights from the 2025 Digital Government Index (DGI) and the Open, Useful and Re-usable Data (OURdata) Index. The DGI analysis comprises 36 OECD countries and 8 accession candidate countries (Argentina, Brazil, Bulgaria, Croatia, Indonesia, Peru, Thailand and Romania), while the OURdata Index presents the same cohort except Denmark and Hungary. The headline results confirm real but uneven progress. The 2025 DGI average rose from 0.61 (out of 1) in 2023 to 0.70 in 2025 This represents a 14% increase, with all but 7 OECD countries improving their scores since 2023 (Figure 1.2). The OURdata Index increased more modestly, from 0.48 to 0.53 (Figure 1.4). Accession countries trail behind: their DGI average of 0.50 is 0.20 points below the OECD average, and their OURdata average of 0.46 is 0.07 lower. These results show that most OECD governments are now closer to the frontier of digital maturity than away from it, but the pace and depth of change remain insufficient to close the gap between institutional capacity and what people and complex policy challenges demand.
Note: 2025 data not available for Germany or the United States. 2025 data cover 1 January 2023 to 31 December 2024. 2023 data not available for Germany, Greece, Slovak Republic, Switzerland or the United States. Data for Indonesia and Thailand cover 1 January 2022 to 31 December 2023. See Annex Table 1.A.1 for detailed country scores.
Source: OECD (2026[2]; 2024[3]; OECD/ADB, 2025[4]).
Progress across the six DGI dimensions is uneven and, when examined through the four transversal facets, reveals a structured pattern. The largest gains were recorded in Data-driven public sector (from 0.63 to 0.74, the biggest improvement of any dimension), User-driven (from 0.61 to 0.71) and Proactiveness (from 0.57 to 0.67). These gains reflect stronger data-governance frameworks, wider use of service design and user testing approaches, and a moderate expansion of the governance and use of AI in government.
Progress has been more limited in Digital by design (from 0.68 to 0.75), Government as a platform (from 0.62 to 0.71) and Open by default (from 0.53 to 0.59). The smaller gains in these three dimensions are partly because countries had already built solid foundations, particularly in governance and cybersecurity, rather than because they have stopped making progress. Key areas such as digital investment governance and monitoring of digital strategies remain weak and have not kept pace with improvements elsewhere in these dimensions.
Open by default results are the most concerning: it recorded the smallest gain of any dimension (from 0.53 to 0.59) and remains the lowest-scoring overall. Progress was driven largely by modest advances in algorithmic transparency guidelines and open-source software policies, while the availability and accessibility of high-value open datasets, the elements that carry the greatest weight in this dimension, improved only marginally. This matters beyond the dimension itself: limited performance in open government data constrains governments' ability to use data for real-time co-ordination, public accountability and innovation across the whole of government.
The four transversal facets that reflect successive stages of the policy cycle (strategic approach, policy levers, implementation and monitoring), and which are tracked across the DGI dimensions, reveal varying degrees of progress. Across all six dimensions, scores for all facets increased compared to 2023 (Annex Table 1.A.2). However, they continue to demonstrate the gap between strategy and implementation (Figure 1.3). Scores for strategic approach (from 0.77 to 0.87) and policy levers (from 0.64 to 0.80) continue to exceed those for implementation (from 0.59 to 0.78) and monitoring (from 0.45 to 0.65) - the stages at which policy intent must translate into operational practice.
Selected transversal facets scores across the six dimensions of the DGI, OECD average, 2023 and 2025
Note: for full list of transversal facets values, see Annex Table 1.A.2.
The results indicate a persistent gap between strategic ambition and implementation, particularly where monitoring lags policy direction. In Data-driven public sector, for instance, the strategic approach score reached 0.80 and policy levers 0.74, while monitoring remained at 0.45, essentially unchanged from 2023. A similar pattern holds in Digital by design, where policy levers reached 0.80 but monitoring remained at 0.55, and in Proactiveness, where despite notable gains in policy levers and implementation, monitoring stayed flat at 0.53. This gap between design and delivery was present in 2023 and has not materially closed, confirming that progress is stronger in establishing formal frameworks and enabling mechanisms than in embedding them into the day-to-day operations, workflows and accountability systems that determine how well digital government performs. Taken together, the evidence shows that the primary barrier to closing the gap is no longer the absence of strategy or policy frameworks, but the failure to translate them consistently into operational practice at scale.
The OURdata results further reinforce this diagnosis (Figure 1.4). OECD countries perform better in Data availability (driven by wider publication of high-value datasets, particularly in education, government finances and justice, the full list of datasets is presented in the annex) and Data accessibility (supported by the EU Open Data Directive requirements) than in Government support for data reuse – the pillar where performance remains weakest. On average, 57% of high-value datasets are available as open data and 67% are accessible through central portals, but only 49% offer application programming interface (API) access. Progress in stakeholder engagement and impact assessment has been modest, confirming that while governments are making data more visible, they are less successful in making it systematically useful for learning, innovation and the creation of public value.
Note: 2025 data not available for Denmark, Hungary, or the United States. 2023 data not available for Hungary or the United States, and not included for Denmark. 2025 data cover 1 January 2023 to 31 December 2024. Data for Indonesia and Thailand cover 1 January 2022 to 31 December 2023.
Source: OECD (2026[2]; 2023[5]; OECD/ADB, 2025[4]).
Taken together, the indices portray a public sector more digitally mature than in 2023, but still insufficiently coherent, adaptive or operationally integrated to close the gaps identified in the previous section. Governments must not only advance individual dimensions of digital government but also ensure that established foundations shape how they operate, co-ordinate and deliver in practice. The next challenge is not only to continue progressing within each dimension, but to close the distance between existing foundations and how governments operate, co-ordinate and deliver in practice. The following sections explore the barriers contributing to this gap and the actions governments can take to translate digital maturity into system‑level performance.
Building on the general and dimension results of the DGI and OURdata Index, this section examines five challenges that explain why the gap between digital ambition and institutional performance persists, and why closing it requires more than adding new components to an already crowded landscape.
1.4.1. Limited data governance prevents unlocking the value of data
Data is central to governments’ ability to act coherently, reduce duplication, anticipate needs and build the conditions for trustworthy AI. The 2025 DGI results show that governments have made genuine progress in treating data as a strategic asset: 94% have data strategies and 92% report metadata standards, reflecting a broad commitment to governing data across the public sector. The use of data for core government functions is also strong with 94% of countries using data for policy monitoring and evaluation, and 92% to anticipate and plan interventions. These are not trivial achievements; they represent a substantial shift from the fragmented, siloed approaches that characterised public sector data management a decade ago.
However, progress has been uneven across the policy cycle. Monitoring of data strategies remain limited, with only 62% of countries reporting that they track results. Data quality management is also incomplete, with only 64% of countries having formal standards for data quality assessment. The gap is sharpest in understanding how services actually perform: just 44% of countries reported having government-wide initiatives to analyse service usage patterns, limiting governments' ability to learn from what people experience and improve services iteratively. Governments will realise limited benefits from stronger data arrangements unless they apply data consistently to service design, delivery and evaluation, not only to planning and oversight.
Open government data shows a similar pattern. Publication requirements and accessibility have advanced: 94% of OECD countries now require public sector institutions to publish open data, and accessibility through central portals is relatively high. But the mechanisms that turn data into public value remain weak. Only 49% of high-value datasets across OECD countries offer API access (a key enabler of reuse), and impact assessment of open data is limited: just 31% of countries evaluate its economic impact, 26% its public sector impact, and only 11% its social impact. Governments are making data visible without making it useful. Without stronger support for reuse and more systemic measurement of outcomes, open government data remain an example of formal availability without practical utility.
1.4.2. Adoption of digital public infrastructure lags behind its rollout
OECD countries have invested substantially in shared digital building blocks - digital identity, data-sharing systems, digital notifications, cloud infrastructure and single digital gateways. The rollout is real, but existence does not automatically produce public value nor integration.
Digital identity governance is largely in place: 92% of countries have a digital identity strategy and nearly all have designated a responsible body. Yet service coverage and actual use remain uneven. 69% report that more than half of their online services can be accessed using digital identity, and population uptake of digital identity ranges from over 90% in nine countries to below 50% in twelve. Furthermore, other systems are highly prevalent across OECD countries, such as data interoperability systems available in 83% of OECD countries, their impact depends on being actively used, which requires human-centred approaches, incentives, compliance and sustained investment that many governments have not yet put in place. Other enabling components such as digital post and digital payments are still absent in many countries.
The result is that many governments have the components for acting as a platform without realising the benefits of doing so. Governments face the greater challenge of driving consistent adoption, integration and effective use across government, as fragmented implementation, uneven uptake and variable interoperability can continue to limit the benefits of shared platforms and data assets. This gap between availability and actual use typically reflects three connected problems: fragmented governance, where no single institution has the mandate or incentive to drive uptake across government; limited requirements for service providers to use shared infrastructure; and insufficient practical support - technical guidance, onboarding support and hands-on assistance - to make common building blocks the default choice. This matters directly for the public. Where digital identity adoption is limited, where interoperability remains thin, and where shared infrastructure is incomplete, governments cannot provide the seamless, joined-up experience people expect and likely struggle to maintain continuity of services when disruptions occur.
1.4.3. Investment, procurement and skills are not yet fully aligned for digital delivery
Governments have strengthened central steering and planning for digital transformation, but the systems that govern investment, procurement and workforce remain better suited to managing compliance than to driving delivery. Ex ante assessment is widespread, with 89% of OECD countries having mechanisms to plan and estimate spending on digital government, but ex-post evaluation is rare: only 25% conduct ex-post cost‑benefit analysis of digital projects. This means governments are approving digital investments without systematically learning whether they work in practice or delivering intended results.
Procurement guidance has improved, with 89% of OECD countries having dedicated central procurement guidelines for digital and ICT projects, but the use of agile and innovation-oriented mechanisms remains moderate; only 55% of countries report using procurement approaches suited to iterative digital delivery. Instead, governments are still struggling to manage fast-changing and data-intensive technologies through static budgets, rigid approval cycles and compliance-oriented controls designed for a different era.
The same rigidity appears in digital talent and skills. Governments are increasingly aware of capability gaps (72% of countries have conducted digital skills-needs assessments) and have expanded training and recruitment efforts. But these remain fragmented and disconnected from long-term workforce planning: only 17% of countries have a dedicated public-sector digital skills strategy. Without sustained investment in people, governments cannot maintain the internal capability needed to exercise sustain strategic control over technology choices, manage vendors effectively or learn from implementation over time. Investment systems that do not learn, and workforces that are not planned, keep governments dependent on external providers and unable to build the institutional memory that sustained digital transformation requires. Together, these barriers limit progress toward more adaptive, evidence-based operating models that support lasting efficiency and resilience.
1.4.4. AI adoption is advancing faster than governments’ capacity to govern it
AI adoption across OECD governments is accelerating, but the safeguards and operational conditions needed to use it safely and at scale are not keeping pace. The strategic foundations are largely in place: nearly all OECD countries have an AI-in-government strategy, 83% have at least one institution responsible for governing AI in the public sector, and AI is now used in at least one area of government in 97% of OECD countries, with strongest uptake in internal processes and public services.
However, the gap between strategy and operational governance is significant. Only 58% of countries provide central support for procuring AI goods and services. Data readiness, interoperability and transparency arrangements, the conditions that make AI trustworthy and scalable, remain underdeveloped in many contexts. Most strikingly, only 28% of OECD countries report conducting any financial or non-financial assessment of the impact of government AI use cases. Governments are adopting AI without systematically evaluating whether it is working as intended. Many have moved from strategy to experimentation, but not from experimentation to durable, government-wide implementation.
This matters beyond AI itself. AI can support productivity, responsiveness and proactivity, but only when built on reliable data, clear accountability and trusted operating arrangements. Where these conditions are absent, AI risks remaining confined to isolated pilots and delivering narrow efficiency gains rather than enabling a genuine shift in government capacity to anticipate, co-ordinate and deliver. The barrier is not ambition or early adoption, it is the limited capacity to govern and deploy AI systems in a way that is trustworthy, measurable and scalable across government.
1.4.5. Services are still too reactive and fragmented to deliver government as one system
Public services are still largely organised around institutional structures rather than people’s needs. Many countries have made real progress: service standards are present in 89% of OECD countries, and co-design tools are widely available. But making standards consequential in practice is harder than establishing them. Formal requirements to apply service standards at both central/federal and sub‑national levels exist in only 44% of countries, and the practical support mechanisms needed to apply them consistently are weaker than the standards themselves. User engagement, while improving, remains insufficiently systematic: the targeted inclusion of groups facing higher access barriers is lower than general co-design activity, and the use of testing methods such as design thinking sessions, focus groups and consultation platforms remain limited.
Measurement of service performance is also weak. Most governments lack the system-wide tools to coherently identify where users drop out, where service journeys break down or where administrative burdens accumulate. Without this feedback, it is difficult to improve services systematically or to demonstrate the value of digital investment to decision-makers and people alike.
The deeper problem is that services remain reactive. People and businesses must still know what to request, when to apply and where to go. The once-only principle, intensive use of data and AI-enabled anticipation, the mechanisms that would allow government to reach out rather than wait to be contacted, remain unevenly implemented. The barrier is not the absence of digital channels but the failure to connect them into a coherent, people-centred system that acts on what government already knows.
1.4.6. Moving away from fragmentation to achieve system-level performance
A common thread runs through all five challenges: governments do not lack individual initiatives, policy instruments or digital components. What they lack is the institutional coherence to make these elements reinforce each other. These include data systems that cannot be shared, infrastructure that is built but not adopted, investments that are approved but not evaluated, AI that is deployed but not governed, and services that are designed but not connected. Each of these is a symptom of the same underlying condition. The barriers are not independent of each other: weak data foundations undermine interoperability; fragmented infrastructure limits service integration; rigid investment systems slow the iteration needed to close capability gaps; and without trust mechanisms, AI cannot scale. The parts are mostly there. The challenge now is making them work as a system, and that requires a different kind of reform effort from the one that built the foundations in the first place.
The five barriers identified in Section 1.4 share a common root: digital government reforms have been more successful at building foundations than at making them work together in practice. Closing the gap between digital ambition and institutional performance is therefore not primarily a question of adding new strategies or components, as most governments already have enough of both. It requires a different kind of effort: connecting what exists, embedding it into how governments operate, and sustaining the organisational conditions needed to learn and adapt over time. The following sub-sections set out four areas where more deliberate and systemic action is needed.
1.5.1. Moving from building digital public infrastructure to driving its use
Critical building blocks like data-sharing systems, digital identity, cloud infrastructure, common registries and service platforms remain underused and insufficiently embedded in daily operations of public institutions. The priority now is more systematic use of what already exists.
This requires designing shared digital components so that reuse is the path of least resistance for users and service providers. When agencies can draw on common digital identity systems, shared data pipelines and standardised service components rather than building their own, they create the conditions for genuinely joined-up government, reduced duplication and more efficient service delivery. Estonia’s approach to digital public infrastructure illustrates what becomes possible when shared infrastructure is actively governed, mandated and maintained over time. Its X-tee data exchange platform, which public sector organisations are legally required to use and which is estimated to save Estonians 844 years of collective work time annually, while digital identity has saved up to 2% of GDP (OECD, 2024[6]).
Reuse does not happen automatically. It requires clear governance including common rules, co-ordination mechanisms and data-sharing agreements that give institutions both the obligation and the practical support to use shared infrastructure. It also requires funding models that cover not just the initial build but the ongoing costs of maintaining, updating and expanding common components over time and across agencies. Infrastructure that is built once and left to atrophy quickly becomes a barrier rather than an enabler.
User trust is an equally important condition for adoption. As governments expand digital identity systems and data-sharing capabilities, users need meaningful agency over how their personal data are processed, used and shared. Finland's Suomi.fi data exchange layer, built on the same architecture as Estonia's system and legally mandated for use by all public sector organisations, combines compulsory adoption with strong data protection safeguards, including clear information about what type of data is shared and for what purpose. This demonstrates that mandating use of shared infrastructure and protecting people’s rights over their personal data are complementary rather than competing objectives. Digital identity systems and wallets that allow users to control and selectively share their credentials, within and across borders, are an important mechanism for building this trust. Where users feel in control, adoption follows; where they are not, even well-designed infrastructure goes unused.
These foundations also determine how far governments can go with AI. Moving AI from isolated pilots to scaled, trusted use depends on reliable data, interoperable systems and repeatable assurance processes. Governments that treat strong digital infrastructure as a precondition for AI deployment, rather than something to be sorted out later, are better placed to realise AI's potential without creating new systemic risks. This need becomes even more pressing as governments harness AI systems to be more proactive.
1.5.2. Making governance and investment fit for digital delivery
The 2025 DGI shows that governance of digital government skews heavily towards planning: strategies are strong, mandates are clearer than before, and ex-ante assessment is widespread. What remains weak is the capacity to iterate, evaluate and adapt once delivery begins. Fixing this requires changes to how governments govern, fund and build capability for digital transformation, not just how they plan it.
Governance models need to embed feedback as a routine function rather than an occasional exercise. This means creating structured mechanisms, including regular reviews, user feedback loops, cross-agency learning forums, that allow strategies to adjust course based on evidence rather than waiting for a formal evaluation at the end of a project cycle. It also means opening governance to external perspectives: governance models and advisory arrangements that bring in industry, civil society and academic expertise help keep digital government strategies grounded in what is technically feasible and socially legitimate.
Institutional structures matter, too. Fragmented central oversight, where digital strategy, delivery support and emerging technology governance sit in separate organisations, weakens the coherence between policy intent and operational practice. The United Kingdom addressed this directly in January 2025 by consolidating its Government Digital Service, Central Digital and Data Office and AI incubator into a single Digital Centre of Government, with a mandate spanning strategy, delivery support and AI governance across departments, an approach designed to reduce the distance between central direction and day-to-day delivery.
Investment approaches need to match the iterative nature of digital work. Large, upfront programme budgets with fixed specifications are poorly suited to technologies that evolve rapidly and whose requirements become clearer through use. Governments that are shifting to modular, staged funding, releasing resources in tranches tied to demonstrated progress rather than upfront commitments, are better able to stop what is not working, scale what works well and avoid locking in costly mistakes. Australia's Digital and ICT Investment Oversight Framework (IOF), which requires major digital projects to pass through staged gateway reviews before receiving further funding, illustrates how investment governance can be structured to promote learning and reduce the risk of large-scale failure. Equally important is closing the evaluation loop: without systematic assessment of whether digital investments have delivered their intended results, governments cannot learn from experience or make the case for continued investment.
Decisions about whether to build, buy or partner for digital capabilities are among the most consequential governments make, yet many governments lack the internal expertise to make these decisions well. Governments that invest in this institutional capability, developing the skills to assess vendor proposals critically, manage contracts actively and understand the long-term implications of technology choices, are less exposed to lock-in, better placed to switch providers when needed and more able to maintain strategic control over their own digital infrastructure.
Ultimately, none of this work without skilled public servants. Expanding training and recruitment is necessary but not sufficient: only 17% of countries have a dedicated public-sector digital skills strategy, and without one, capability-building efforts remain fragmented and reactive. Korea's digital government talent development programme, part of the country’s Human Resources Development System, illustrates how workforce planning and digital delivery can be systematically connected by combining a dedicated skills framework with structured career pathways for digital roles across the public sector. Governments need workforce plans that develop the cross-disciplinary skills, combining policy, service design, data and technology, that digital delivery requires, and that create career pathways attractive enough to retain talent over time. Public servants who understand both the policy intent and the technical realities of digital systems are the most important enabler of the shift from digital foundations to digital performance. Many governments expand training and recruitment but do so outside a long-term workforce plan.
1.5.3. Building trust into AI in government from the outset
The OECD Trust Survey shows that people do not yet sufficiently trust governments’ handling of emerging technologies, AI or personal data. The 2025 DGI offers some justification for these concerns: transparency, algorithmic accountability and impact assessment remain among the weakest areas of digital government performance across OECD countries.
This challenge is intensifying as AI becomes more deeply embedded in government. Conventional AI tools that generate content or recommendations already raise questions of accountability and bias. Agentic AI systems – which can place and execute sequences of actions across government tools and workflows, on behalf of users or public servants – raise the stakes further (Box 1.2). They offer the prospect of more proactive, integrated and responsive public services, but they also may also raise the stakes regarding clear oversight mechanisms, traceable decision trails, meaningful accountability, traceability and user control and the ability to pause, reverse or challenge automated actions. Governments that take account of these considerations from the outset will be better positioned to scale agentic AI responsibly.
AI agents are systems that can perceive and act upon their environment with a degree of autonomy, using tools as needed to achieve specific goals and adapt to changing inputs and contexts. Agentic AI generally refers to systems composed of multiple co-ordinated AI agents that can break down tasks, collaborate and pursue more complex objectives over time with limited human intervention. While generative AI (GenAI) systems answer, agentic AI systems act.
This distinction is important for government because the implications shift from what an AI system says to what it does. Conventional machine learning (ML) or GenAI tools may support analysis, drafting or interaction. By contrast, agentic AI systems may retrieve information across databases, trigger workflow steps, route cases, initiate transactions or update records within defined permissions. This creates opportunities to reduce administrative burden, improve responsiveness and simplify services, but it also raises the bar for legality, accountability, traceability and human oversight. Governments therefore need to govern not only the model, but also the action space granted to the system, the approval points in the workflow, and the mechanisms for oversight and redress.
An early public sector example is Estonia’s Bürokratt, an early-stage interoperable network of government AI solutions that, from the user’s perspective, functions as a single channel and unified gateway to government services. In the envisioned (and early) user experience, a user can request a complex service such as ordering a new passport, and an AI agent orchestrates the interaction across the relevant authority. The broader aim is to move service complexity into the background through system-to-system interaction while allowing the user to validate key actions where needed.
The OECD Working Party of Senior Digital Government Officials (E-Leaders) is exploring how governments can use agentic AI responsibly, including its public sector use cases, readiness requirements, and the enablers, guardrails and engagement needed for trustworthy adoption.
Source: (OECD, 2026[7]; OECD, forthcoming[8]).
Transparency is a foundational element for deploying and using AI in government. Governments should proactively disclose information about the algorithms they deploy, including their purpose, the data they use and the safeguards in place to prevent bias or error. The Netherlands' Algorithm & AI Register, developed in direct response to public concerns over the use of a faulty algorithm by the Dutch Tax Office and now covering over 600 algorithms across nearly 200 public organisations, offers a recent and well-documented model for institutionalising algorithmic transparency at scale. Publishing algorithm registers, strengthening open data practices and records gives users meaningful visibility and control over the use of their personal data are concrete steps that reinforce public confidence. The same principle applies to how governments govern and share personal data more broadly: informing people on the use of their data, and developing digital identity systems and wallets that give users genuine control over their credentials, rather than simply digitising existing processes, are a practical expression of transparent design.
Alongside transparency, independent assurance builds confidence that digital transformation is safe, fair and working as intended. This means embedding ethical frameworks and risk management processes throughout the lifecycle of digital systems – not only at the point of deployment – and establishing oversight mechanisms with the independence and technical capacity to hold governments to account. Modular and staged approaches to digital delivery support this: by breaking large programmes into smaller, evaluable components, governments can demonstrate progress, surface problems early and maintain accountability throughout. The 28% of OECD countries that currently assess the impact of government AI use cases represent a floor, not a ceiling. Systematic impact assessment should become a standard expectation, not an exception.
1.5.4. Delivering public services as one coherent system
Governments have invested heavily in individual service improvements. The next step is to connect these into service journeys that work across institutional boundaries, so that people experience government as a single, coherent system rather than a series of separate organisations each requiring its own interaction. This is less a technical challenge than a governance one: it requires deliberate choices about how services are designed, how data flows between institutions and how performance is measured and held to account.
The foundations for this shift are increasingly in place. Digital identity, data interoperability and the once-only principle – whereby governments collect information from users once and reuse it across services rather than asking for it repeatedly – make genuinely integrated, multi-agency service journeys technically feasible at scale. Denmark's digital public infrastructure, including MitID and NemLogin digital identity ecosystem and data-sharing systems, have enabled a range of life-event based services, covering everything from birth registration to business start-up, that pull together information across agencies without requiring users to re-submit data they have already provided. What is needed now is the governance and operational discipline to apply them consistently. Service standards, co-design methods and user testing are increasingly widespread, but the next step is to embed them in everyday delivery, decision‑making and improvement routines. Strengthening these mechanisms through clearer mandates, better guidance and more systematic measurement of user experience is essential to turning isolated service improvements into system-wide change.
AI, and agentic AI in particular, offers significant potential to accelerate this shift. In a public-sector context, agentic AI can help navigate complex, multi-agency processes on user’s behalf, identifying relevant entitlements, assembling information across systems, completing routine steps within defined limits and prompting users ahead of key deadlines or life events. Over time, this could move much of the complexity of interacting with government into the background, making services feel more proactive, connected and responsive. But realising this potential depends on the same foundations that underpins integrated service delivery more broadly: interoperable systems, trusted digital identity, clear records of automated actions, and meaningful points at which user can review, challenge or override what the system has done. AI amplifies the value of strong foundations, and the risk of weak ones.
Services also need to be designed for continuity and cross-border use. Continuity means maintaining service access when circumstances change or demand surges, through multi-channel delivery, fallback mechanisms and the ability for users to move between channels without losing progress or repeating steps. Cross-border operation means allowing people, businesses and goods that move across borders to access services in other jurisdictions as seamlessly as they do at home. This requires interoperable digital identity systems, trusted data-sharing frameworks and mutual recognition of digital credentials. Both are expressions of the same principle: services designed around people's actual situations, not around the organisational convenience of individual institutions.
The analysis in this chapter presents the central challenge the rest of this Outlook addresses: governments have built real digital foundations, but have not yet translated them into consistent, system-level performance. The OECD Digital Government Outlook provides a comprehensive, forward-looking assessment of digital government policies across 36 OECD Members and 8 accession candidate countries. Drawing on the results of the 2025 OECD Digital Government Index and the Open, Useful and Re-usable Data (OURdata) Index, it evaluates progress and persistent gaps across key areas of digital transformation, identifying what governments need to do to move from digital ambition to public sector performance in an environment of rapid technological change, fiscal constraints and limited public trust.
The Outlook is structured around this overview chapter and four thematic chapters covering key areas of digital government policy:
Chapter 2 examines how OECD countries build and use digital infrastructure and data as foundations for more effective and future-ready government. It reviews progress and gaps in rolling out digital public infrastructure, expanding the use of digital identity, and implementing data strategies and data-management practices.
Chapter 3 explores how countries develop the governance and organisational capacity needed to manage digital transformation. It examines developments and challenges in steering digital government, managing investments, and building the digital talent and skills to prepare the public workforce needed to deliver it.
Chapter 4 presents how countries govern and adopt AI in government. Following the report Governing with AI, it provides a quantitative overview of the enablers governments have put in place to ensure trustworthy use, and progress towards more user-centred and responsive AI-enabled services.
Chapter 5 examines how countries are building human-centred and proactive public administrative services in the digital age. It explores how governments organise, steer and improve services so that people experience government as reliable, joined-up and easy to navigate.
The cross-country analysis presented in this Outlook is accompanied by country notes for each participating OECD Member and accession candidate country, setting out the state of play and identifying strengths and areas for improvement on critical aspects of digital government policy.
Note: 2025 data is not available for Germany and the United States. It covers the period from 1 January 2023 to 31 December 2024.
Source: (OECD, 2026[2]).
OECD average, 2023 and 2025
Note: The transversal facets assess performance across the stages of the public policy cycle.
Source: (OECD, 2026[2]; OECD, 2024[3]).
Note: 2025 data is not available for Denmark, Hungary, and the United States. It covers the period from 1 January 2023 to 31 December 2024.
Source: (OECD, 2026[2]).
Note: 2025 data is not available for Denmark, Hungary, and the United States. It covers the period from 1 January 2023 to 31 December 2024. Sub-pillar names are Content of the open by default policy (1.1), Stakeholder engagement for data release (1.2), Implementation (availability of high-value datasets) (1.3), Content of the free and open access to data policy (2.1), Stakeholder engagement for data quality and completeness (2.2), Implementation (accessibility of high-value datasets) (2.3), Data promotion initiatives and partnerships (3.1), Data literacy programmes in government (3.2), Monitoring impact (3.3).
Source: (OECD, 2026[2]).
Note: The categories of high value datasets are determined by the OECD and primarily based on the G8 Open Data Charter. Datasets are only considered available if they are free of charge, machine-readable and provided with an open license, following the OECD definition of open data from the Recommendation on Enhancing Access to and Sharing of Data.
Source: (OECD, 2023[5]).
References
[2] OECD (2026), “Digital Government Index and Open, Useful and Re-usable Data Index: 2025 Results and Key Findings”, OECD Working Papers on Public Governance, No. 90, OECD Publishing, Paris, https://doi.org/10.1787/6347ec74-en.
[7] OECD (2026), “The agentic AI landscape and its conceptual foundations”, OECD Artificial Intelligence Papers, No. 56, OECD Publishing, Paris, https://doi.org/10.1787/396cf758-en.
[3] OECD (2024), “2023 OECD Digital Government Index: Results and key findings”, OECD Public Governance Policy Papers, No. 44, OECD Publishing, Paris, https://doi.org/10.1787/1a89ed5e-en.
[6] OECD (2024), “Digital public infrastructure for digital governments”, OECD Public Governance Policy Papers, No. 68, OECD Publishing, Paris, https://doi.org/10.1787/ff525dc8-en.
[5] OECD (2023), “2023 OECD Open, Useful and Re-usable data (OURdata) Index: Results and key findings”, OECD Public Governance Policy Papers, No. 43, OECD Publishing, Paris, https://doi.org/10.1787/a37f51c3-en.
[1] OECD (2020), “The OECD Digital Government Policy Framework: Six dimensions of a digital government”, OECD Public Governance Policy Papers, No. 2, OECD Publishing, Paris, https://doi.org/10.1787/f64fed2a-en.
[8] OECD (forthcoming), Governing with [Agentic] Artificial Intelligence, OECD Publishing, Paris.
[4] OECD/ADB (2025), Government at a Glance: Southeast Asia 2025, OECD Publishing, Paris, https://doi.org/10.1787/bc89cb32-en.