Why the Apprenticeship Levy could be key to AI
- AI success depends on investing in skills, not cutting training funding.
- AI apprenticeships help close critical skills gaps.
- Protecting the levy is key to the UK's AI ambitions.
Like most of my colleagues, I was heartened by the ambition and scale of the UK government’s AI Action Plan. I also think it can work. However, it requires people and skills – data engineers, machine learning specialists, security professionals – that are currently in shortage. Half the challenge will be to find and train the people to make this happen.
The good news is there is already a solution out there. There are vast sums of available money right now that can help address this shortage. The bad news? Not enough businesses seem to know about it.
AI, as I’m sure you’ve read already, is a great disruptor. It will likely render some jobs unnecessary – but create totally new ones. Other roles will be forced to change and adapt. Upskilling, reskilling, and growing new talent will be the way forward. It’s clear then, that apprenticeships should be a pivotal part of the discussion.
For those who know how apprentices work in the UK, organisations with a wage bill over £3 million contribute 0.5 per of their payroll costs to a pot of funds that can then be used for apprenticeship training. This is called the Apprenticeship Levy.
Those companies are then eligible to claim £15,000 to offset against any apprentices they take on. And if you think all apprentices are carpentry or plumbing – think again. A growing proportion is about technology training, and uptake is consistently on the rise. Across the tech and digital apprenticeships we offer, we have seen a 27% increase in Starts (apprentices enrolled onto programmes) across the market, over the past two years.
However, those available funds are currently woefully underused. As of today, an appalling £2 billion worth of unused Apprenticeship Levy funding has been returned to the Treasury under the existing ‘use it or lose it’ rules, between its launch in April 2017 up to 2021 alone.
Each year, an average of £500,000,000 is returned to the treasury in unspent levy. That's enough to train 25,000 AI Data Engineers a year, one of the most in-demand careers globally.
Two billion pounds of, effectively, free training money, has been handed back to the government rather than spent on training people. And at a time when we face a well documented skills shortage.
If we want the AI Action Plan to really work – this has to change.
We can get training funds into every business, with levy transfer
There are two issues: Awareness and access.
Many leaders don’t fully understand their levy funds are even there, or what to use them for. Working with a training provider is critical to spend your pot wisely.
One big problem is that the levy is only accessible to those large organisations; seemingly counterintuitive when 99.9% of private sector businesses in the UK are small to medium enterprises (SMEs). It’s the law of small numbers - we need all businesses equipped to embrace AI, not just the minority of bigger dogs.
As it stands, larger companies, through something called the levy transfer, first introduced in 2019, can pledge 25% of their unused funds to businesses in their supply chain and support the growth of both business and talent.
But not enough of them are doing it. Two billion pounds has gone back to the Treasury already – a lot of that could have gone to smaller companies. Actually using the levy transfer could turn the tide on wasted training funds, and put us into high gear on the journey to becoming an AI superpower.
The reason not enough large companies are utilising levy share is: it’s too complicated! Conversations with their supply chain organisations are disjointed, and the onus is on them to execute a plan. The government can relieve this pressure; simplify matters to drive levy share uptake and reduce wastage. It’s about more straightforward rules and support from the organisations now driving the UK’s ambition to be an AI superpower. The Government’s AI Hubs initiative should have access to unused levy funds and drive effective redeployment nationally.
It’s not just funding, but also learnings that need to filter down. Plenty of small businesses are willing but not ready; they’ve drunk the Kool-Aid on AI, but no one is telling them where to begin. A roadmap based on advanced learnings (as well as a boost of funds) from the larger enterprises making major inroads can help move the dial on a national scale.
Why transfer your levy funds?
Why are bigger organisations not taking up the opportunity to pledge levy transfer, and how can we encourage them?
Firstly, awareness of levy transfer is insufficient. Communication from the top down from government can move that needle. Someone, somewhere should be screaming from the rooftops that there are hundreds of millions of pounds worth of free training available. I’m sure most SMEs don’t even realise it’s there. Perhaps if they did, they would start lobbying the Levy contributors to help them out – especially if they are part of that company’s supply chain.
In terms of the benefits, it boils down to social responsibility. When it comes to our future as a nation, we get to decide now whether we will fall on the right side of what my colleague Vicky Crocket has coined the ‘AI divide’. We can do it by skilling at scale, and sharing access with a wider talent pool. It’s not a ‘nice to have’, but a new imperative.
The truth at present is that there’s not enough talent in data science, data engineering and machine learning. Even large enterprises struggle to get good data talent, such is the shortage. This leaves SMEs unable to compete.
The talent that the AI Action Plan demands will be built from the ground up. That can be reskilling and upskilling within businesses, or SME’s hiring local talent to train and grow. A new generation of professionals, kickstarted by the government’s AI regions initiative (a system of geographic ‘AI hubs’ focused on finding people, and building skills sustainably for the future) and fostered by businesses putting their skills to work – if they’ve got the levy funds.
So, what is in it for you, big enterprise? Why would you just hand all that free training over?
It’s about the long game. You can be part of building a stronger data-driven AI-powered future. Or, you can find yourself in a different future altogether, where in five, or ten years there is still insufficient data talent for us to realise the ambitions that the UK government is wholeheartedly behind.
Don’t let unused funds be the thing that stops us. What a waste that would be.
Yes, we can become an AI superpower
Sir Keir Starmer wants the UK to become a global AI superpower. There are a few questions that follow… Can we really achieve it? How soon? And how?
For me, the answer is a resounding yes to the first. The UK is well-positioned and well-equipped. We already punch above our weight in terms of research; with the tools and skills to start doing, we can move fast.
In line with the Labour Government’s ambition to deliver this new reality within the current term, I think the action plan can succeed in as little as three years.
That depends of course on the ‘how’; yes, we must build compute power, new infrastructure, and data centres on another scale – but doing so means nothing if we don’t enable the population at large to put the technology to work. Human intelligence will be the key to meaningful progress.
Interested in taking advantage of the levy fund for AI apprenticeships? Find out more about the levy fund and our AI apprenticeship programmes.
Related articles
Deidré Luzmore explains why effective AI governance starts with leadership, accountability, and clear organisational oversight.
AI governance can't wait for regulation. Leaders can't either
6 November 2026
Cyber Portfolio Director Richard Beck shares practical insights on AI governance, assurance, accountability, and the risks of autonomous AI agents.
Expert spotlight: Richard Beck on AI governance, assurance, and accountability
6 November 2026
As AI moves from experimentation to widespread adoption, many organisations are discovering that governance is struggling to keep pace. While attention often focuses on future regulations, today's challenge is ensuring employees, leaders, and processes are aligned around responsible AI use.
As AI adoption accelerates, governance lags behind
4 November 2026
An unauthorised AI collective, a real-world cyber intrusion, and the uncomfortable question of whether human oversight and AI governance can keep pace.
When AI agents started talking to each other
26 October 2026
As AI becomes more capable and autonomous, who has the evidence, independence, and authority to decide when the risk is no longer acceptable?
Who decides when AI is too dangerous?
24 September 2026
David Pool explains how to build a cost-efficient AI enterprise by balancing model, governance, data, FinOps, and workforce capability.
Building the cost-efficient AI enterprise
10 September 2026
Why healthcare organisations must prioritise workforce transformation alongside AI adoption to achieve sustainable and effective change.
Before AI transformation comes workforce transformation
8 September 2026
Anthropic has announced a research preview of its Model Hardware Standard (MHS), a new specification designed to help AI agents interact with physical devices. If Model Context Protocol (MCP) gave AI a common way to discover and use software tools, MHS does something similar for hardware.
Anthropic's Model Hardware Standard: what happens when AI agents can use physical tools?
28 August 2026
Forward deployed engineers are emerging as a key AI role, helping organisations bridge the gap between technical capability and business value.
What the rise of the forward deployed engineer tells us about how organisations are building capability
27 July 2026
London’s plans to prepare workers for AI-driven change reflect what QA’s analysis of AI readiness across UK sectors has found: adoption is accelerating, but workforce skills, confidence, and capability aren’t keeping pace. By examining how AI is being adopted across industries, we’ve identified a widening gap between giving people access to AI and equipping them to use it effectively.
Closing the AI readiness gap starts with skills
21 July 2026
About the Author
David Pool