Around three-quarters of businesses investing in AI are not getting the return they hoped for, according to a recent IBM survey of chief executives. The reason is rarely the technology itself – it is people: the leaders and teams who do not yet have the skills, or the confidence, to put AI to work. I hear this from employers almost every day. At the same time, the apprenticeship budget is set to reach £3.3 billion in 2026-27. For all that potential investment, conversations about the levy too often feel like a chore. Spend it, transfer it, or lose it. At the same time, the government’s Employer Skills Survey showed that the percentage of employers offering training has dropped from 66 per cent in 2017 to 59 per cent in 2024. The opportunity is to reverse that decline by using the levy to build the capability businesses actually need. No organisation can hire its way to AI transformation. Success depends on building capability across the whole workforce, from the senior people who set strategy and sign off the risk, to the teams putting AI tools to use day to day. The employers already pulling ahead are the ones aligning their levy spend with their business priorities. They treat AI as a leadership and people challenge rather than just a technology project, and they invest accordingly. This is what turns a promising commitment to AI adoption into lasting productivity, and one-off training into productivity that sticks. This is why the emerging reforms to the growth and skills levy matter. One of the most significant is the arrival of apprenticeship units: shorter, focused modules that give employers more flexibility and targeted development in priority areas. Three of the first seven units focus on AI leadership, covering strategy, governance, procurement, and organisational transformation. This is good news. The people who set an organisation’s risk appetite and sign off its AI investment are often the least confident about the technology, and the least likely to commit to a full apprenticeship. A focused unit gives them an accessible, convenient way in and avoids employers being encouraged to use apprenticeships where is too long and doesn’t align with the learner’s job role or their ability to complete the programme. Units will not fix everything though, and we should be honest about that. They will add to the skills system, not replace full apprenticeships. Critically, they must not become a backdoor to thinner, cheaper, lower-quality training. The apprenticeship quality requirements built up over the past decade exist for good reason, and they need to stay. The real value of a unit is as a gateway: a way for employers to engage with structured, quality training, and to see for themselves what deeper investment might deliver. Their success will depend on two things above all. The first is how quickly the model evolves. As employers, learners, and providers start using units in practice, we need to gather feedback in real time and adapt, because AI is moving far too fast for the skills system to stand still. The second is whether employers help to shape units rather than simply taking products off the shelf – they need to be genuinely employer-led, not formulated by well-meaning civil servants or ministers. The strongest apprenticeship standards have always been the ones employers built, and the same will be true here. Units are only one part of a skills system that is evolving at every stage. From Alan Milburn’s review of young people and work, to the emergence of V Levels and the Francis review’s recommendations for schools and colleges, through to ensuring the growth and skills levy continues to support adults already in work, change is happening across the entire landscape. The real opportunity lies in joining these reforms together rather than treating them in isolation. Only then can the system deliver more than the sum of its parts. The real prize in the growth and skills levy is not simply greater flexibility. It is the chance to reposition skills development as what it truly is: an opportunity, not a compliance chore. The organisations that thrive in the AI era will not be the ones that spend the most on technology. They will be the ones that invest most effectively in their people. Apprenticeship units are the latest chance to do exactly that, and we should embrace them, and help to shape them.