Turning AI ambition into public value

The AI-enabled transformation of organisations in all sectors is progressing at pace, with a view to realising the opportunities it offers. The need to integrate AI into government activities and processes is now widely recognised, whether to counteract a shortage of skilled workers, operate more efficiently, or make state services digitally accessible. However, the outcomes remain uneven. Most public sector AI implementations are being approached and managed in a bottom-up manner, resulting in a disparate multitude of AI solutions that typically offer only limited benefits. While many public sector organisations have now launched promising AI use cases, they often struggle to scale them. 

Overall, the highest levels of maturity were found in Governance. By contrast, implementation of Systems & Technology is lagging behind, and development of data skills (People) is not yet a priority for public sector organisations in most countries. Taken together, this indicates that while strategic intent may exist, execution needs to improve. Some European countries stand out for consistently taking a more integrated and strategic approach to data governance, infrastructure, and workforce development. These countries—our 'Europe front-runners'—are Switzerland, Poland, Norway, Estonia and the Netherlands. This group is closely followed by countries in the Middle East & Asia, which also score relatively highly across the three pillars. North America shows moderate maturity in all areas. When front-runners are excluded, Europe ranks lowest across all three pillars, indicating systemic challenges in aligning governance structures, technology frameworks, and skills-building initiatives. Globally, countries that perform well in Governance also tend to score higher in Systems & Technology and People. This implies that strong governance frameworks support superior technological infrastructure and skills development.

Key insights across the pillars

Our survey highlights compelling findings across each of the pillars. Explore these in more detail and below and download the PDF for further insights. 

Larger countries tend to face greater challenges in achieving high data quality—with more organisations involved, there will be a wider range of stakeholders to coordinate and governance structures are likely to be more complex. Such challenges are likely to be especially significant in federally organised countries. At a central level, governments in these countries will have high data and AI maturity. But at the federal and local level, there will be proounced variations in approach that often make it impossible to monitor and enforce central guidelines.

In-depth analysis

Even when the region's front-runners are included, Europe consistently scores lower across nearly all Governance dimensions, including data quality measurement and accountability. This indicates challenges in coordination, strategic maturity, and enforcement and oversight mechanisms. By contrast, countries in the Middle East & Asia and North America regions generally perform better, suggesting more centralised and/or more effective governance structures in those regions. Countries outside Europe rate themselves significantly higher for Governance, with stronger foundational governance structures than their European counterparts. Some of the core building-blocks for success include 'hub & spoke' governance models, centralised reporting with KPIs spanning data quality, centralised incident response structures, and enforcement of standards by cross-ministerial decision-making bodies.

What good looks like

Canada

Canada’s data strategy is a pioneering example of data governance, as it establishes centralised oversight  for the entire public sector under the auspices of the Treasury Board of Canada Secretariat (TBS). ​ Binding standards are defined through the Office of the Chief Information Officer (OCIO), and compliance with these standards is audited annually via the Management Accountability Framework (MAF). ​ This process measures the compliance of ministries using precise indicators, such as the maintenance of up-to-date data inventories, the implementation of metadata standards and the appointment of responsible data stewards. ​In this way, Canada transforms strategic guidelines via the TBS into a measurable and highly binding control mechanism for the entire federal administration. ​

Operationalisation lags behind strategy. Country rankings are consistently lower in Systems & Technology than in governance practices. Legacy systems, federalism, and metadata misalignment are the primary implementation barriers. A significant central–local government divide persists, with local administrations often lacking the resources to operationalise common standards. This creates fragmented systems that complicate data aggregation and AI value creation. Bottom-up initiatives outperform top-down approaches. Grassroots efforts (e.g., Germany's Data Labs) deliver tangible results, while top-down programmes often stall due to red tape or competing interests. Overall, these findings underscore the need to close the strategy–implementation gap, requiring multi-level, tailored approaches that account for the structural realities of each country.

In-depth analysis

Metadata and data catalogue maturity is highest in the Middle East & Asia, with Europe lagging behind. Centralised data catalogues and data-sharing mechanisms are more advanced in North America and Middle East & Asia, indicating strong cross-entity coordination and interoperability in those regions. Enterprise data models and architecture frameworks appear to be challenging for government organisations worldwide, with low maturity scores across all regions. AI use-case management and lifecycle governance are well structured in Middle East & Asia, while processes for deploying and monitoring AI models appear to be more difficult in Europe. Development of use cases is uniformly low across all regions, indicating a shared need for more structured collaboration and innovation processes.

What good looks like

Germany 

Data Labs were initiated through the 2021 Federal Data Strategy, which outlined plans to establish a data lab in every federal ministry. These decentralised, ministry-specific innovation hubs enable data exploration, analytics, and AI experimentation directly within each department – rather than relying on a single central authority. Several ministries have already established their labs, with rollout continuing across all federal ministries. Examples of data labs that have already been established:​

• Federal Foreign Office (AA) - This lab operates PLAIN, a cross-ministerial platform for secure data-sharing and collaborative analysis. It also develops AI-based tools that help diplomats analyse and compare policy documents to improve foreign-policy work.​

• Federal Ministry of the Interior (BMI) - The BMI data lab works closely with the Federal Statistical Office, providing analytical support, rapid data assessments, and AI-related services. ​

• Federal Ministry for Economic Cooperation and Development (BMZ) - The BMZ's data lab strengthens data and AI capabilities in international development cooperation, shaping data frameworks and fostering responsible data use for sustainable development.​ ​

The data labs created a new institutional framework for data-driven policymaking. Since 2025, the newly established Federal Ministry for Digital Affairs and State Modernisation(BMDS) coordinates this approach, harmonising standards, governance, and best practices across all labs.​

Skills efforts are growing but still early-stage. Governments increasingly recognise the need for data and AI skills, yet programmes are rarely AI-specific and lack monitoring mechanisms to assess progress. Governance maturity drives skills maturity and countries with stronger governance structures also tend to perform better in workforce development. Horizontal use cases dominate, with broadly applicable applications (e.g., document processing, chatbots, translation) widely adopted due to low barriers and cross-departmental relevance. High-impact vertical use cases remain rare, as domain-specific solutions (e.g., healthcare analytics, tax fraud detection) are held back by limited specialised skills and resistance to organisational change. Overall, realising the full potential of data and AI skills development will require governments to move beyond generic training programmes toward targeted, AI-specific upskilling, and to invest in the specialised skills and change management needed to unlock high-impact vertical applications across key public service domains. 

In-depth analysis

Middle East & Asia indicate more structured efforts in building and monitoring data and AI skills, supported by structured national programmes and strategic investments. North America follows, showing moderate progress, while Europe's positioning suggests slower advancement and less emphasis on upskilling the workforce in Data-and-AI-related competencies. These findings highlight the need for more targeted and measurable approaches to AI skills development, especially in regions where governance and infrastructure are still evolving. Best-in-class countries distinguish themselves through a strategic and structured approach to data and AI skills development and centralised monitoring mechanisms. ​Learning is guided by national standards, supported by dedicated institutions, and tracked through regular reporting. Leading countries align skills development with broader governance and infrastructure efforts, creating a coherent framework for workforce transformation.

What good looks like

The Netherlands 

The Dutch central government has established a comprehensive, role-based strategy for developing data and AI skills across all ministries. Through the national IT Academy (RADIO), civil servants follow structured learning paths that begin with foundational data literacy and the widely accessible ‘National AI Course’, and progress to advanced topics such as machine learning, ethics, and domain-specific applications. ​

Core competencies—including digital literacy, data literacy, and AI awareness—are centrally defined and monitored via the Kwaliteitsraamwerk IV (KWIV). Ministries report annually on staff skills, training hours, and workforce flows, while departments tailor specialised curricula to their policy domains. This ensures a shared national baseline and fosters deep, context-specific expertise.

Our research points to four foundational success factors for these mission-critical initiatives: ​

1. Leadership commitment and data strategy: ​ Data structure and quality, accessible, well-managed data are critical enablers for scalable AI. A top-down commitment to organising and prioritising data governance is essential, with a cross-ministerial decision-making body that can define, enforce, and monitor common standards across all departments. At the same time, decentralised implementation units within each ministry must be empowered to act, adapting central standards to their specific needs and ensuring consistent execution across the public sector.
2. Investing in resources and capabilities: Prioritise investments that strengthen data and AI capabilities by allocating more financial, technical, and human resources, including upskilling staff, modernising infrastructures, and building dedicated teams. To justify continued investment, organisations should focus on making the benefits of data and AI measurable — developing frameworks to track return on investment (ROI), monitoring outcomes, and communicating value to stakeholders
3. Standardised data catalogue: This is essential for creating transparency and consistency in public sector data usage. By establishing a unified catalogue enriched with metadata, and aligning it with a common tools landscape, public administrations can promote shared standards and guidelines across departments. This foundation enables scalable and interoperable data practices, allowing decentralised entities to work efficiently while adhering to central requirements — a key enabler for cross-government collaboration and the effective deployment of AI solutions.
4. ‘Hub & Spoke’ / grassroots lab model: Grassroots integration of data and AI labs is a key success factor. Rather than starting with top-down regulation or centralized planning, these labs focus on rapidly implementing use cases to demonstrate value and feasibility, allowing teams to experiment, learn, and iterate before aligning with broader strategic frameworks. Setting up such initiatives centrally would take significantly longer due to the complexity of standardisation and coordination. Instead, the lab model—seen in countries such as Germany—enables direct implementation first, followed by the development of best practices and guidance that can be scaled and shared across government entities

We conducted a study based on a structured questionnaire completed by PwC specialists working in public sector and AI-related roles across 19 countries worldwide to assess the maturity of data and AI capabilities in their country's public administration across three pillars —Governance, Systems & Technology, and People.

Maturity ratings were based on our proprietary data governance maturity model, which shows that in order to leverage their data assets effectively, governments must set central standards and build up a comprehensive target model to enforce application of these rules. At the same time, government entities need sufficient leeway to tailor their data assets to their specific business needs. 

Our model differentiates between three different levels of maturity: ​

Decentralised: This is the model typically followed in the early stages of data maturity. There are no centralised data standards, and government entities have independent or siloed data agendas ​

Centralised: This is the interim phase, where a central data authority provides and enforces policies, standards and guidelines, while also taking on an execution role on priority projects, such as integration of address data. ​

Federated: Here the centre retains policies, standards and guidelines for development, while individual entities take on capability-specific governance and operational roles (including through centres of excellence), drive execution, and provide bottom-up input to policy. ​

This end-to-end maturity journey is divided into five stages in each of the three key pillars.