We Have Been Teaching Entrepreneurship Wrong for Decades

The consensus in education circles is that we need more entrepreneurship programmes. More pitch competitions, more enterprise days, more young people being told they could be the next big thing. The intent is right. The approach, I think, is not.

Because we have been running entrepreneurship education for decades. And the results are telling. Research from Capital on Tap found that 82% of 18 to 24-year-olds in the UK are considering starting their own business. That is an extraordinary level of ambition. And yet only 4% manage to establish sustainable enterprises on their first attempt. The ambition is there. The execution is not. And we keep designing programmes as though motivation is the problem we need to solve.

The gap nobody wants to name

The problem with how we teach entrepreneurship is that we have always treated the doing as someone else's problem. We give young people the inspiration and hand them the ambition. Then we expect the world to meet them halfway. It does not.

Execution requires resources. It requires time, money, networks, and - most critically - the capacity to build things. Historically, that capacity cost a lot. You needed a team of ten to do what one person with clear vision could only dream about. Low-income young people, as a scoping review of 46 UK studies makes clear, face the steepest barriers: financial, social, and human capital all working against them at once. The execution gap falls hardest on those with the least existing support. And our education system has largely accepted this as given.

Here is what I think people are not seeing

AI changes the economics of execution. That is the frame I keep coming back to. The cost of doing - of building, testing, iterating - is dropping toward zero in ways that are genuinely new. A teenager with a clear idea and access to AI tools can now build what previously required a team, a budget, and three years of experience. That is not hype. That is a structural shift in what is possible.

What that means is the case for entrepreneurship education at genuine scale has never been stronger. Not programmes that teach young people about entrepreneurship. Programmes that help young people execute. There is a real difference. The first produces more pitch decks. The second produces more businesses.

The GEM UK National Report from September 2025 found that 36% of UK working-age adults are now engaged in or planning to start a business - the highest level since records began in 1999. But the majority remain at the intention stage, not the execution stage. That is the gap AI is positioned to close. Not by doing the thinking for people, but by removing the barriers that make doing feel impossible.

What will young people do for work in an AI world?

The honest answer is: build things. The assumption that young people need to choose between getting a job and starting a business was always a false binary - but now the cost of the latter has collapsed enough that it becomes a genuine option for many more people. AI lowers the barrier to execution, which means entrepreneurship stops being a privilege reserved for those with existing capital and networks. The real question is whether our education system moves fast enough to meet that shift.

What this means in practice

If you are responsible for learning and development or employability programmes, the implication is straightforward. The old model - teach concepts, hope for inspiration, leave execution to chance - will not produce different results. The new model treats AI as infrastructure, not as a topic. Young people do not need more lessons about entrepreneurship. They need environments where they can build things, with tools that make doing accessible.

That is what Sherpas AI is designed to be. Not a course about entrepreneurship. A system that helps young people execute.

I might be wrong about the timeline

AI capability is moving quickly, and the educational infrastructure to support this kind of learning is still being built. But if I am right in broad terms - that the cost of doing is falling fast enough to make execution genuinely accessible to young people for the first time - then the failure to equip them with AI tools is no longer a resource constraint. It is a choice.

If you are thinking about what AI-era learning looks like for the people you develop, the learning and development director page is a good place to start.