The OECD’s 2026 work on AI and skills argues that capability has to be connected to the conditions of work: access to learning, the chance to practice, and a productive role in which new skills can be used. It points toward a dual response: widen access to high-quality learning and improve the match between skills and productive, rewarding work.
The usual response is to buy training. Training matters, but it cannot repair a system that makes new capability difficult to use. A person can complete an AI course and still return to a workflow with unclear decision rights, poor data, no time to practice, and incentives that reward the old behavior.
Capability has four layers
We help leaders distinguish four different questions:
- Knowledge: Does the person understand the method, tool, or policy?
- Practice: Does the workflow give them a safe place to use it repeatedly?
- Authority: Do they have the decision rights to act on what they know?
- Reinforcement: Do measures, managers, and systems make the behavior durable?
Most transformation plans over-invest in the first layer and under-design the other three. That is how organizations end up with a highly trained workforce still waiting for approvals, navigating duplicate systems, or escalating decisions nobody owns. The OECD’s emphasis on skills governance and skills matching suggests a broader conclusion: capability is an access-and-deployment problem, not just an education problem.
Design the role before designing the curriculum
For an enterprise moving into AI-enabled work, the right unit of change is often a role or service, not a course catalog. Start with the work: what should the person notice, decide, delegate, review, and learn? Then decide what the technology can support and what the management system must protect.
For a public agency, this may mean pairing a new analytical capability with explicit review and documentation responsibilities. For a healthcare system, it may mean redesigning the team huddle around a new data signal. For a technology organization, it may mean giving platform teams the product and communication skills to turn infrastructure into a service.
The Global Enterprise view
Our leadership and talent work connects capability maps to operating models, change architecture, and the future of the work. The objective is not simply to make people ready for change; it is to make the organization easier to learn inside.