AI Isn’t a Tech Skill: What Somerset’s Skills Plan Gets Right (and Wrong) About the Future of Work

Yesterday, I was interviewed for Somerset’s Local Skills Improvement Plan. Not because I love plans with logos and deadlines, but because the LSIP is one of the few levers we have that can genuinely influence how our region prepares people, businesses, and educators for what work looks like next.

Not in a vague, buzzwordy way, either. In the very practical sense of: “What are employers actually struggling with, what do people really need to learn, and how do we stop training from lagging two years behind reality?”

The interview was with Shelley, whose day job is Future Skills Lead at Somerset Council. Her mission is refreshingly simple: get as many voices as possible into the plan, so it actually reflects Somerset rather than sounding like it was written in a vacuum.

I was happy to contribute. If we want a skills system that actually works, we need employers being honest about what’s missing, training providers being brave enough to evolve provision quickly, and policymakers giving programmes the stability to actually function long-term.

That last bit? We’ll come back to it.

What the LSIP Actually Does

In plain English, the LSIP is government-funded and designed to feed directly into post-16 education and training decisions. The aim is that colleges, training providers, and other decision-makers can adapt provision based on what employers are saying they need, not what someone assumes they need from a spreadsheet.

It’s a good idea. Whether it works depends entirely on whether anyone actually listens to what comes out of it.

Why I Keep Saying This: AI Is Not a “Tech Skill”

If you only take one idea from this article, make it this:

AI is not really a tech skill.

Yes, it’s a technology. Yes, it emerged from the tech world. But using it well is mostly about communication, delegation, critical thinking, ethics, and the ability to learn, adapt, and iterate without needing someone to hold your hand.

Those are life skills.

That’s why AI is such a fascinating (and contentious) topic. It forces people to confront how they think, how they work, and how they collaborate. One of the more persistent myths we encounter is the assumption that “tech people” automatically understand AI better. They don’t. It landed on everyone at roughly the same time. In fact, tech backgrounds can sometimes get in the way, because people overcomplicate what is, fundamentally, a communication tool.

At Techosaurus, we’ve now worked with 150+ businesses through the Generative AI Skills Bootcamp for Somerset. We’ve seen accountants run circles around software developers. We’ve watched people with zero tech background build automations that would make a startup jealous. We’ve also seen highly technical people freeze up because they’re waiting for “the right way” to prompt something, when there isn’t one.

The pattern is clear: curiosity beats credentials every time.

The Skills That Will Matter Most (and the Ones I Actually Look For)

When we bring people into our orbit as associates, I’m looking for experience more than a narrow list of “skills.” Skills can be taught. Curiosity, judgement, and real-world perspective are harder to manufacture.

The traits I value most are:

Curiosity, because you can’t teach it, and the future belongs to the curious.

Critical thinking, because AI will happily produce plausible nonsense if you let it.

Clear communication, because prompts are persuasion, not programming.

Confidence to experiment, because so much learning is “What does this button do?” followed by reflection on the outcome.

Delegation mindset, because AI works best when you treat it like a junior teammate. You still need to guide it, check it, and refine it.

I can teach someone the tech pieces. The harder part is giving them the judgement that comes from actually being in business, managing people, dealing with customers, and shipping work under pressure. That’s not something you get from a qualification. It comes from doing the work.

Why We Refuse to Deliver Bootcamps Online

This one polarises people, but I’ll stand by it.

Our bootcamps are face-to-face because they’re 60 hours of learning and discussion. Online learning is convenient, and it can be brilliant in the right context, but you cannot build a genuinely safe space for messy, nuanced conversations when half the room is multitasking, checking notifications, or replying to emails.

And we need that safe space, because we talk about ethics, bias, whether using AI is “cheating”, data privacy, real-world workplace use, and the fear people feel about being left behind.

You don’t get honest contributions if people don’t feel present. You also don’t get the group dynamics that make learning stick.

One of the best surprises from the bootcamp has been the community that formed as a by-product. We now have an alumni group, we run events, and we’ve seen people who had never met before end up sharing tips, tools, and opportunities purely because they’re united by curiosity.

I genuinely didn’t set out to build that community. I wanted to teach. The community happened anyway, and now, if I’m honest, it’s a huge part of why I love doing this.

From Generative AI to Automation: Brains and Hands

We’re also evolving what we teach, because AI is moving fast. But the bigger shift isn’t “new features”—it’s how it changes workflows.

The framing we use is simple: AI is the brains, automation is the hands.

It’s one thing to generate content, images, or ideas. It’s another to set up systems that actually do things for you. An automation layer can read an incoming email, recognise that it’s finance-related, and route it to the right team without you touching it. The benefit isn’t “wow, look at the tech.” The benefit is time, focus, and fewer interruptions.

People rarely realise how much mental overhead they’re carrying until it’s removed.

That’s why we’re now launching our Automation Skills Bootcamp, building directly on the foundations of generative AI. I’m genuinely excited to see what our learners go on to create, because the combination of AI thinking and automation doing is where the real productivity gains live.

The Real Limit on Growth Isn’t Demand—It’s Certainty

A question that comes up a lot is capacity for growth.

We’ve invested, we’ve built a small associate model, and we can deliver more. We also have a nice flywheel where we sometimes spot talent on the bootcamp and then invite those people to work with us. But that only works when you’re delivering regularly and you can genuinely offer opportunities.

The biggest limiter is funding certainty.

If a programme relies on government-backed funding, it becomes a gamble whether that funding will exist in six months. I’m not a gambling man. I don’t want to “grow on a whisper,” hire people, and then have nothing for them to deliver. That’s not business, that’s roulette, and it risks livelihoods.

We’re diversifying into private work as well, because many organisations want hands-on support. But for regional skills impact, bootcamps are powerful when funded properly and communicated clearly. The current model feels like trying to build a house on shifting ground.

If policymakers want training providers to invest in capacity, they need to commit to more than 12-month cycles. You can’t scale expertise on uncertainty.

What I Want Colleges and Providers to Stop Doing With AI

I said this in the interview, and I’ll say it again here:

Please stop bolting AI into the “digital” bucket and leaving it there.

When AI sits only with digital teams, it stays in a tech echo chamber. It gets taught in tech language. It gets framed as “a tool you use later,” rather than “a capability you apply everywhere.”

AI fits everywhere.

I used the example that you can teach geography without AI, or geography with AI. In an underfunded, overpopulated classroom, AI can help differentiate learning for 30 to 40 students in ways that a single teacher physically cannot. A student can ask for a different explanation, a new analogy, or extra practice, and the AI can tailor it on demand.

That’s not replacing teachers. It’s giving teachers leverage. But it needs guardrails and good pedagogy to avoid turning into a “copy and paste” shortcut.

The analogy I like is this: AI is the calculator for human language. We don’t ban calculators because someone could “cheat.” We teach maths properly, and then we teach how to use tools responsibly.

The same principle applies here. Embed it across the curriculum. Teach verification alongside generation. Treat it as a literacy, not a specialism.

The Skills Gap No One’s Talking About

Two areas stood out strongly in the conversation, and neither of them are about AI:

Real-world finance literacy. So many people start businesses without understanding basics like turnover versus net profit, VAT nuances, and how to make practical financial decisions. Our convoluted laws and regulations don’t help, but this isn’t taught clearly enough early on, and it has a massive impact on whether a new business survives.

Soft skills, especially among younger entrants. This one is delicate, but I’ve seen it repeatedly. The online world has changed how people communicate. Picking up the phone can feel like conflict. Disagreement can feel like rejection. Yet healthy conflict is part of work, and learning how to navigate it is vital.

AI will not fix that. If anything, AI raises the bar: the more automation you have, the more valuable human judgement, empathy, and communication become. You can’t automate your way out of difficult conversations. You can only get better at having them.

Real-World Experience Works, Especially for Students Who’ve Been Told “No” Too Often

One of the most uplifting examples we discussed was a pilot scheme I had the privilege of being involved in, hosted at iAero here in Yeovil. Local Year 10 students who hadn’t been able to secure traditional work experience placements got to collaborate on real projects, using AI, to design new shoes with all the marketing and presentation work that goes with such a project. Clarks funded a lot of it, and the winning designs were actually manufactured as real shoes.

A student later gave a talk saying he would never have done that before, but the programme changed how he saw himself.

That story matters because it shows what happens when you give people a real challenge, modern tools, supportive facilitation, and a tangible outcome. We should be fanning that spark. The system should be designed to create more moments like that, not fewer.

The Mentorship Gap

This is slightly adjacent to skills, but it matters a lot for outcomes.

As a founder, time is the bottleneck. There are business support groups, coaches, and programmes, but not all advice is created equal. Anyone can call themselves a coach. That means founders can waste time and money following generic guidance that doesn’t apply to the real constraints of running a business in Somerset.

What I’d love to see more of locally is credible mentorship programmes, founder-to-founder guidance, and clear pathways for “what next” after early traction. If we want more local employers to grow and create jobs, we need to treat mentorship as part of the skills ecosystem. It often gets left out because it’s hard to measure, but that doesn’t make it less important.

Practical Takeaways for Somerset’s Skills Future

If you’re an employer, educator, policymaker, or someone who cares about how the region develops, here’s what actually needs to happen:

For employers: Treat AI adoption as a business capability, not an IT project. Protect time for real CPD, not just “here’s a link” training. Recruit for curiosity, judgement, and communication, then teach the tools. The tools will change every six months anyway.

For colleges and training providers: Embed AI into every subject area where language and reasoning matter—which is almost all of them. Teach AI alongside ethics, critical thinking, and verification habits, not as an isolated module. Build more project-based, real-world experiences, especially for students who struggle to access traditional placements, because those students are often the ones with the most untapped potential.

For policymakers and LSIP stakeholders: Stabilise what works. Bootcamps can be powerful, but only if funding and communications aren’t a constant cliff edge. Invest in mentorship pathways for founders and small employers, because growth capacity is a regional skills strategy in disguise. Stop describing AI as a purely “digital skill.” It’s a productivity skill, a communication skill, and a leadership skill too.

Closing Thoughts

I came away from the interview feeling hopeful, because the opportunity in front of Somerset is real.

If we can align businesses, educators, and decision-makers around a simple truth—that AI is a cross-cutting capability rather than a niche tech topic—we can help a lot of people get better at what they do, reclaim time, and build more resilient careers.

My prediction: within two years, AI literacy will be as expected as email literacy. The organisations that treat it as a shared capability will move faster, adapt better, and do so with less drama. The ones that outsource it entirely to IT will be playing catch-up, wondering why their competitors seem to get more done with fewer people.

If the LSIP can help push us toward that first group, then this conversation—and the dozens like it happening across Somerset right now—will have been worth it.

Because skills planning isn’t about predicting the future. It’s about making sure people have what they need to build it themselves.