AI capability does not turn into AI value on its own. 80% of companies have deployed generative AI in at least one function, and 80% of those see no meaningful effect on earnings. When electricity first reached factories, output barely moved until businesses redesigned their workflows and retrained their workers.
For AI to produce value, two things have to happen at once: human capital has to adapt to AI, and AI has to adapt to human capital.
Human capital adapts through learning and better-run labour markets. Our education system was designed for a world that no longer exists, so we need graduates who are adaptive and ready to work with AI and manage it on higher-level tasks: that is our Learn thesis. The labour market is poor at moving the right people into the right roles, and AI will multiply job transitions, so we need one that places people well and helps them navigate their careers: that is our Navigate thesis.
AI adapts when it is built for the work people actually do. Left generic, its gains land only at digital-first companies and the workers who are already winning, while most workers and most of GDP sit in health, education, the public sector, construction and manufacturing. AI built for those industries, meshed with their data and physical reality, is what keeps us from a two-speed economy: that is our Work thesis.