The Level 5 AI Leadership Apprenticeship has a fundamental problem: it appears to be measuring success by contact hours delivered per pound, not by outcomes.

That might seem like a minor detail about funding metrics, but it’s created a requirement for 90 hours of classroom time across the three modules—more than two full working weeks—that locks out the people who need this training most: SME leaders.

I’ve spent 15 years delivering technical training to SMEs. I helped build the Level 4 Software Developer Apprenticeship. I’ve delivered Skills Bootcamps in AI to SME leaders. And the single clearest thing we’ve learned is this: classroom training doesn’t work for this sort of learning. Practice does.

The Level 5 curriculum looks solid—it covers exactly what SME leaders need to understand about AI adoption and digital transformation.

But the delivery model makes it nearly impossible for us to deliver, and difficult for SMEs to access.

Because when you optimise for contact hours per pound instead of productivity gains, you end up mandating 30 hours of classroom time per module—90 hours total if you do all three.

That’s more than two full working weeks outside of work. And that creates a barrier that most SME leaders simply can’t cross.

Who Was in the Room?

I don’t know who Skills England consulted when designing this. My guess is they talked to large organisations—the sorts of places with training budgets and people who can afford to be out of the business for days at a time.

What’s missing is input from people delivering training to SMEs. People who’ve tried to get an SME director into a classroom for 30 hours and learned what works and what doesn’t.

When we delivered AI Skills Bootcamps aimed at SME leaders, the biggest blocker wasn’t cost. It was time.

The funding was there. The content was relevant. The need was clear. But SME leaders couldn’t afford to be out of their business for extended periods. Not because they didn’t want to learn, but because their businesses can’t function without them for that long.

According to the OECD’s 2023 D4SME Survey, 43% of businesses dissatisfied with their digital adoption cite “lack of time for training” as the main cause. Even more striking: 47% of those businesses didn’t participate in any training activities at all.

That’s a design problem, not a motivation problem.

The 90-Hour Problem

The Level 5 apprenticeship mandates 30 hours of classroom contact time per module—90 hours total across all three—on top of the general off-the-job training requirements for apprenticeships.

For an SME director, that’s not just a significant ask. It’s a blocker.

You’re asking someone already doing three jobs—operations, sales, and strategy—to block out more than two full working weeks for classroom sessions. Sessions that don’t align with their specific business problems. Sessions that pull them away from the context where they need to apply what they’re learning.

The Skills Bootcamps we delivered required 60 guided learning hours over 12 weeks. That meant one or two sessions a week—and even that was too much for most SME leaders. The time outside the business was the main barrier stopping people from accepting the course.

The ones who did it saw real gains. But we kept coming up against the same problem: freeing people up to make that first step.

The Level 5 apprenticeship takes that barrier and multiplies it. Ninety hours of mandated classroom time is a non-starter for most SME leaders.

What works is spreading it out—one or two sessions a month over three to four months, with supported practice in between. That gives people time to apply what they’re learning, see the gains, and reinvest that saved time into the next session.

The hurdle isn’t the total time. It’s freeing them up to make that first step. Once they’ve done the first few sessions and started seeing productivity gains, they’ve got that time back to invest in the rest.

But you can’t front-load 90 hours of classroom time and expect SME leaders to make it work.

Why Practice-Based Learning Works

We started working on what became the Level 4 Software Developer Apprenticeship back in 2012. It began as a volunteer project—a group of us trying to create a curriculum where people could actually learn software development properly, in practice.

One thing became clear almost immediately: classroom training doesn’t work for learning these sorts of digital skills.

We had to do everything as practice and supported practice. That insight has shaped every programme we’ve designed since—including the Level 4 AI and Automation Apprenticeship I designed in 2025.

When I started designing that Level 4 AI programme, I began with practice. What are we going to practise over these 12 months? What’s the journey? Then I reverse-engineered the qualification from understanding that journey in practice.

Our model is built around that. Short group sessions to introduce a topic or method, then learners practise it at work, solving actual business problems. They’re supported through digital tools—Slack, for example—by their cohort and expert trainers. Then they come back for reflective coaching sessions to review what they’ve done and how to improve.

Most of the learning happens solving real business problems, not sitting in a room listening to someone talk. You need support in practice, not mandated classroom time.

The Real Barriers to AI Adoption

UK SMEs aren’t struggling with AI adoption because they don’t understand the technology. The technology is actually simple compared to the complicated languages and systems we’re used to teaching people. The real barrier is that SME leaders are overwhelmed with everything else.

The UK Government’s SME Digital Adoption Taskforce found that despite the UK’s world-class tech sector, UK SMEs lag behind G7 competitors on the uptake of digital tools and AI. Many overrate their readiness, hampered by time, cash, and know-how gaps.

And according to the British Chambers of Commerce, 43% of SMEs have no plans to adopt AI at all.

That’s not a knowledge gap. That’s a time, context, and capacity gap.

Just saying “the training’s there, why don’t you take it” isn’t enough. We need to help SMEs actually be able to do this. Providing training they can’t access isn’t okay.

SME leaders need to see how AI solves their specific problems. They need to try it in their own context. They need support when it doesn’t work first time. And they need accountability to keep going when other priorities pile up.

Ninety hours of classroom time doesn’t deliver that. It just creates another barrier.

What Good AI Leadership Training Looks Like

We’ve been through this process ourselves. Recently, we automated a commission payment process that used to take two hours every month. It now runs completely automated, with just two hours once a month to check it.

That took 45 minutes of work to set up.

But it didn’t happen because someone sat in a classroom learning about automation frameworks. It happened because we’d built a culture of continuous improvement, peer learning, reflection, and feedback over time.

You can tell someone about psychological safety in a classroom. But putting it into practice takes time, dedication, support, and—most importantly—practice.

To put something into practice takes practice. You do things, you make them better, you do them again. You need a support network—your cohort provides that. You need guidance—technical experts provide that. You need reflection and accountability—the peer network and structured check-ins provide that.

That’s what makes it stick. If you try something and it doesn’t work out, who keeps you accountable for continuing? What improvements have you made this week? If that doesn’t exist in your workplace, it needs to exist in the support structure.

The Level 5 apprenticeship structure makes this method difficult to deliver.

What This Means for SMEs

The programme has launched. We can’t deliver it the way it’s been structured—we won’t get the results with the mandated contact hours, and the funding level makes it difficult to deliver anything of real value.

So we’ve built direct-sale courses instead.

Other providers might find ways to make it work. But for us, the structure doesn’t align with what we know works for SMEs.

The consequence is that SMEs end up paying for training that should be funded. They’ll look for alternatives, and the market for quick AI training is already flooded with options that vary wildly in quality.

Meanwhile, the UK government’s own taskforce has calculated that just a 1% productivity uplift through digital adoption would add £94 billion to GDP annually.

The SMEs we’re working with through our direct-sale courses are seeing immediate gains. But I really wish there was government funding for this. SMEs are having to pay for training that should be funded—if it were designed to fit how they actually work.

What a Better Framework Would Look Like

When you look at the skills, knowledge, and behaviours in the Level 5 apprenticeship module, it’s solid. I agree with what’s in there. The problem is how it’s set up to be delivered.

Here’s what I think would work better:

Treat it like a small, fast apprenticeship. Not broken into three separate components. One module where we do a skills gap analysis, work out what people actually need, and deliver it accordingly over three to four months. The key is off-the-job training—that’s what makes apprenticeships work. It lets us deliver supported practice in the business, on real business problems, with time to apply, reflect, and improve between sessions. Mandated contact hours don’t give us that flexibility.

Fund it properly. Good training in this area costs money. Expert trainers aren’t cheap. But it’s not just trainer costs—it’s onboarding, compliance, support, making sure it actually works for each business. That takes people time. There’s some irony in that: we’re selling efficiency through AI and automation, but delivering it well requires human attention. Which is exactly the point. We use automation to free up time for the people-to-people work that actually matters. Right now, the funding level is too low to deliver that properly.

Allow flexibility and alignment to business needs. Every SME is different. The AI tools they need, the processes they want to improve, the time they have available—it all varies. Rigid contact hour requirements make it impossible to tailor training to what each business actually needs. Flexibility is what makes this work in practice.

We already deliver all our apprenticeships this way. We’re set up for skills gap analysis, flexible delivery, co-creating training plans for different employers and different people. The infrastructure exists.

If SMEs could access the funding and we could deliver this properly, we’d be well on the way to unlocking that £94bn opportunity. Instead of the current 1% adoption rate, we’d have SME leaders equipped to lead AI transformation in their businesses—government-funded, practical, and aligned to real needs.

I wrote an open letter to the government last year about making apprenticeships work better for SMEs. This is part of that same challenge: getting the design right so the funding actually reaches the businesses that need it most.