From Ambition to Action

Governments and educators face a huge mission: prepare youth and adult learners for the jobs of the future—a future that will unquestionably be shaped by the rapid development and adoption of emerging technologies such as AI. But the biggest constraint is not technology. A lack of institutional alignment across ministries, education providers, employers, and funders is slowing the translation of AI strategy into learning and economic outcomes at scale.

That is the central finding from BCG’s work supporting the inaugural Bett Ministerial Symposium in January 2026, where we discussed the evolution of teaching and learning in the age of AI with education ministers, policy influencers, and edtech innovators from 54 countries. Our agenda focused on a pair of foundational questions: what will drive successful digital transformation in education and workforce development, and how can we address and avoid the most common barriers to effective implementation?

The private sector has made its position on the future of AI in the workplace clear, as corporations plan to double their AI spending this year and elevate AI strategy to a top C-suite priority. With shifting labor market requirements and heightened expectations, historically slower-to-move education systems are now under pressure to act on a similar scale.

This report explores the challenges facing these systems and the solutions playing out around the world. The case study sidebars that appear throughout this report consist of information and insights drawn from our interviews, from BCG research, and from online sources. We also highlight some direct quotes in the pages that follow.

The stakes are high. Education is the primary long-term enabler of countries’ economic, social, and political development. However, technological changes are outpacing institutions' capacity to respond.

What separates the leaders from the rest is disciplined progress on four fronts: setting a clear, industry-connected ambition; investing in people; dismantling structural barriers to adoption; and scaling innovation beyond pilots.

What We Heard at Bett—and Worldwide

Countries vary significantly in their approach to both the “what” and the “how”: what they prepare learners to do and how they alter practices around AI and tech-enabled learning.

Everyone is doing something to address teaching methods—the “how”—such as by pushing out new tools to educators, changing how teachers assess students, and funding experimentation. But the consistency and boldness of these efforts are far from uniform. Notably, an education system’s engagement with new technologies reflects the country’s culture, governance, and attitude toward digital adoption.

We also observed that countries at the forefront of innovation in tech education are clearer than others about what they teach and why. They connect teaching principles to national ambitions, such as the aspiration to be a global leader in quantum computing or information technology.

On the basis of our conversations at Bett and our work with education systems worldwide, we have defined four broad stages of AI-in-education maturity:

• Foundational. Systems are investing in basic digital infrastructure, such as connectivity, devices, and introductory AI literacy. Policy frameworks are nascent. Experimentation is limited and largely uncoordinated.
• Experimenting. Pilots are underway across classrooms, districts, or institutions. Early results are promising but isolated. Systems are still defining governance and funding models. Scaling remains elusive.
• Scaling. National frameworks are in place. Educator capability-building is systematic. Evidence of impact is accumulating, though consistency across regions and learner groups remains a challenge.
• Systemic. Systems have embedded AI across the full learning life cycle, from early education through workforce development, and they link funding models to outcomes. Public-private alignment is strong. The system learns and adapts continuously.

Most countries represented at Bett sit in the foundational or early experimenting stages. Only a few—including Estonia, Singapore, Kazakhstan, and South Korea—are making credible progress toward scaling. The case studies that follow illustrate what meaningful progress looks like at different stages, not as a linear path, but as a set of hard-won advances that others can learn from.

What to Teach About Tech Transformation

As technology changes, durable human skills become more critical. A clear consensus emerged across the countries represented at the Bett Ministerial Symposium: the most future-ready learners are not those with the deepest technical skills, but those that combine foundational AI literacy with distinctly human capabilities that technology cannot replicate, such as critical thinking, resilience, communication, curiosity, and ethical judgment. The goal is to equip students with the skill sets and mindsets they will need in a human-led, AI-amplified world, where new technologies emerge constantly.

Education needs to focus on the application of skills. The private sector has an essential role in curriculum design, not as a downstream consumer of graduates, but as an upstream co-designer of the skills pipeline. (See “Connecting Instruction to Industry.”) Working together, stakeholders across the education ecosystem can identify the competencies that industry and society will need, create structured opportunities for learners to apply them, and incorporate mechanisms to anticipate how demand will evolve. Although education leaders are eager to activate public-private partnerships, the engagement of employers and technology partners in the design process remains uneven.

How to Teach Technology at Scale

AI adoption is more than a tool rollout. At the Bett UK 2026 edtech exposition, we encountered a vast market of available tools and applications: productivity ecosystems, assessment platforms, classroom workflow modules, assistive technology, and immersive learning environments. The abundance is both exciting and overwhelming. Practitioners consistently called for clearer guidance on what works at what level of implementation risk in their specific context.

Systems will have to adapt end-to-end. Pedagogy is just one dynamic component in this AI era, and systems that treat it as the only one will fall short. Assessment practices, governance frameworks, learner accessibility, educator capability, and data infrastructure all need to evolve in parallel, within a framework where safety and accessibility are non-negotiable design principles. National leaders may consider not just how planners will build top-down frameworks, but how educators who are stretched, skeptical, or under-resourced will implement them. (See “Redesigning Education End-to-End for the AI Era: South Korea.”)

Putting AI in the Right National Context

Where a country sits on the AI education maturity spectrum and how it progresses are inseparable from its governance model, cultural context, and digital history. There is no single right path.

The starting point matters less than the direction of travel. Countries come to the AI table with different priorities, approaches to guiding citizens, and levels of digital maturity. Some nations are investing to upgrade foundational technical infrastructure in schools; others are already at the forefront of digital innovation and platform interoperability. Some favor consistency in implementation and prescribe technologies for educators to use; others focus on offering a guiding framework and encouraging individual experimentation.

But whatever their context may be, leaders told us that inaction is not an option. Even as they navigate different cultural milieus, they encourage swift decision making to keep pace with the relentless progress of technology.

Evidence building is the compass. With such rapid change, longitudinal evidence regarding which innovations actually work is still emerging—and leaders, who have been flying blind on this crucial question, are hungry for information. Stakeholders told us that they want to invest in rapid-cycle measurement to drive decision making. In addition, they want to find new ways to fund fit-for-purpose investments and curate global communities of practice to share implementation nuances.

Different governance structures—from centralized systems to teacher-led experimentation—shape how countries pursue AI-enabled education. (See “Using Context to Drive Strategy: New Brunswick (Canada) and France.”)

The Challenges Facing Stakeholders

The opportunity presented by AI education is real, but so are the structural headwinds. Over half of all employees globally will need to reskill or upskill by 2027. Many skills expire faster than they can be replaced: the half-life of technology-related skills has shrunk from roughly five years to two and a half. The education system faces a shortage of 44 million teachers by 2030, even as public budgets tighten. And BCG research shows that most organizations still lag in establishing responsible AI guardrails, posing a risk that increases when the users are students and the stakes are developmental.

The leaders we spoke with describe a near-universal set of challenges in this area. (See the exhibit.) Such obstacles are not excuses for inaction; instead, they call for planners to devise bold implementation strategies to overcome specific barriers.

The Systems That Inspire Us

Both the barriers and progress against them are well documented. Across the countries represented at Bett, leaders are addressing impediments to AI-enabled education in four distinct ways: setting a clear ambition, investing in people, dismantling structural barriers to adoption, and innovating at scale. No country has perfected any of these approaches, but the systems that inspire us are moving deliberately, learning from early missteps, and building toward durable success.

Setting a Clear Ambition: The “Why” of the Work

Rolling out AI tools is a necessary but insufficient step. Countries on the path to durable progress start by asking, “What kind of a future are we preparing learners for, and what does that demand of our education system today?” Answering those questions clearly, publicly, and in alignment with national economic priorities distinguishes systems that are undergoing genuine transformation from those that are still finding their footing.

Setting ambition is not an abstract exercise. It entails defining outcomes, documenting talent and skill gaps, building multiyear roadmaps, and working toward alignment across ministers, government departments, and private partners. (See “Aligning Vision with Government Priorities: Saudi Arabia and California (US).”) In the private sector, transformation hinges on C-suite commitment. In the public sector, the equivalent element is coherence across funding, procurement, implementation, and governance functions. Achieving that coherence is hard and rarely complete, but systems that actively pursue it move faster than those that do not.

Investing in People: The “Who” of the Work

AI-enabled transformation is primarily a people and change management challenge. This is clear in the business world, where BCG has found that leading organizations consistently follow the 10-20-70 principle: they focus 10% of time and energy on the algorithm; 20% on data and technology; and 70% on people, processes, and cultural transformation.

The education sector is no different. AI generates a lot of emotions—excitement, reluctance, caution, urgency—and both learners and teachers need to be convinced of the technology’s value to them. The 2026 Digital Education Outlook published by the Organization for Economic Cooperation and Development (OECD) notes that, despite positive data points, educators are fearful that AI could undermine their autonomy and compromise safe learning. And even though many students have fully jumped onto the usage bandwagon, researchers have found notable differences in adoption across different ages and geographies—as well as shared concerns about the accuracy, reliability, and ethical implications of AI.

A significant proportion of students say that they feel underequipped for an AI-enabled workplace, and a relatively small percentage of teachers feel confident about their ability to guide students. The constraint is not primarily a matter of technical training; it extends to limits in motivation, belief, and institutional support. BCG's experience is that educators who understand why AI matters to their students' futures and who feel genuinely supported in experimenting are far more effective than those who receive tool training alone. (See “Prioritizing People: São Paulo (Brazil) and Singapore.”)

Dismantling Barriers to Adoption: Freeing Infrastructure, Funding, and Policy

Overcoming policy-, budget-, and infrastructure-related barriers to AI education takes ambition and deliberate action.

Infrastructure is a foundational need—but stakeholders can’t wait for it to catch up. A harsh reality faces stakeholders: without reliable infrastructure, efforts to teach AI can falter. Many markets have to deal with problems involving connectivity, energy demands, and device availability. Locking in investments in devices, data centers, cloud, or energy supplies will remain an ongoing agenda point for policymakers. 
 
Waiting for perfect conditions before committing to take action is itself a strategic choice, and it can be a costly one. Ukraine is advancing its STEM-oriented AI curriculum despite being engaged in active military conflict, as it prepares its workers to rebuild the country’s physical infrastructure. Several African countries, including Rwanda, Kenya, Senegal, and Uganda, are expanding and pairing school connectivity with digital learning platforms—some of them zero-rated, offline-capable, or mobile-accessible—to reach underserved learners. (See “Pushing Beyond Infrastructure Limits: Ghana.”)

AI innovation needs funding innovation; old models won’t work. Many education systems still think of technology as a one-time purchase rather than as a long-term investment. But a comprehensive assessment of the full life-cycle costs of AI would layer in infrastructure, data architecture, teacher upskilling, maintenance, cybersecurity, and continuous platform upgrades on top of the core technology.

Cost transparency is necessary, but systems could also use smarter funding strategies tied to intended outcomes. At Bett, leaders described mounting fiscal pressure due to combinations of shrinking public investment in education, problematic debt servicing costs, and declining development aid. Against this backdrop, policymakers and multilateral organizations such as UNESCO are promoting exploration of new models, including results-based financing, income-contingent loans, and consortium purchasing.

Funding transformation requires more than fiscal creativity. It calls for a shift in the way systems define and measure success. (See “Why Old Ways of Financing Won’t Work.”) Leaders who can articulate clear outcomes, build data systems to track them, and structure partnerships around shared accountability will unlock financing models that input-based budgeting never could. The technical instruments exist. The limiting factor is institutional readiness to use them.

Policy ought to protect, not encumber. At Bett, we heard a call for targeted policies to address AI usage, with guardrails around data privacy, classroom integrity, and fairness. Trust is paramount. The goal is for students to engage in learning without taking intellectual shortcuts or being subjected to inappropriate oversight or profiling. When appropriately managed, access to and engagement with technology will not reinforce bias with regard to racial, socioeconomic, gender, or disability-based groups.

Multilateral organizations, including OECD, UNESCO, and UNICEF, play a key role in shaping AI policy, bridging the gap between rapid technological change and slower-to-move national regulation. For example, UNESCO launched its Recommendation on the Ethics of Artificial Intelligence in 2021, and since then 193 member states have adopted it as the global standard for AI governance. The Recommendation emphasizes human rights, transparency, accountability, fairness, and data protection and allows innovation to uphold public values.

Meanwhile, countries and regions are taking steps to craft their own policies, with input from local stakeholders. (See “Putting Policy into Action: Estonia.”)

Innovating at Scale: From Pilots to Critical Mass

Pilots are easy; scaling is hard. Education systems today are peppered with islands of innovation that deliver compelling results at the classroom, district, or state level but rarely translate into systemic change. BCG has found that although over half of the organizations it surveyed report experimenting with AI in limited areas, only 8% are doing so at scale. The public sector lags even more, despite facing some of the highest perceived disruption risk from generative AI.

The countries that are making the most progress in this area share a common orientation: they treat scale not as the final stage of innovation, but as a design principle from the start. (See “Scaling Success Stories: Kazakhstan and the US.”) They identify barriers specific to their context and digital maturity—connectivity, funding, policy climate, educator readiness—and build implementation architectures that naturally accommodate growth. Critically, they do not wait for certainty. They combine rapid iteration with clear guardrails, they treat early failures as learning cycles, and they move faster than the technology itself.

Bringing It All Together

Dozens of conversations later, our perspective on what drives successful digital transformation in education and workforce development has evolved.

Ambition is not what separates education systems that are transforming from those that are just beginning their journey. Nearly every minister we spoke with has a vision. Instead, the critical difference is disciplined translation of that vision into aligned action across institutions, funding streams, and people.

The levers that we have described are interdependent initiatives, which means that progress on each reinforces the others. Systems that invest in people without introducing structural reforms produce capable educators working inside capacity-constrained institutions. Those that pursue structural reforms without adopting a scaling mindset produce pilots that fail to reach more beneficiaries. The levers rarely move in perfect sequence, but consistently neglecting any of them undermines the rest.

For leaders who want to determine where to focus next, the diagnostic questions are straightforward: where are you on the AI-in-education maturity curve, and at which stage does your system's progress currently stall? The answer will point directly to the lever that demands priority attention.

The governments and education systems that are making the most progress share one further trait. They have stopped waiting for certainty before acting. There is no single model for AI-enabled education, but hesitation is a guaranteed path to failure.

Moving now is a big part of the answer to Bett’s key questions—what will drive successful digital transformation in education and workforce development, and how can we address and avoid the most common barriers to effective implementation? The window for decisive action is open. We are eager to see the global education community document progress, expand bright spots, and share emerging solutions in the months to come.

The authors thank Anna Bryant, Addie Chamberlain, Faisal Faraz, Kevin Fu, Amanpreet Grewal, and Dara Kovacheva of BCG and Ed Clark, Emily Colyer, Alexandra Constantine, and Anna Wood of Bett for their valuable contributions to this report.