From prompt engineering to soft skills: The new ‘skills-first’ baseline for EU entry-level roles

Early career hiring at scale comes down to one thing most Fortune 1,000 employers get wrong: they optimize the screening process for credentials instead of learning velocity.

A student from a non-target school who built something real — shipped code, ran an experiment, broke something and fixed it — will outperform a stronger-credential candidate who has only done coursework. The problem is that credential screening is easier to systematize, so it wins by default.

What works: give candidates a small, ambiguous problem with no right answer and watch how they approach it. Not a LeetCode exercise — something that requires them to ask clarifying questions, make a judgment call, and explain their reasoning. That predicts performance far better than GPA or university ranking.

For government agencies specifically: the compensation gap with private sector is real but overstated for early career. Mission clarity and structured growth paths matter more to new graduates than most agencies assume. The agencies that retain early talent are the ones that give real responsibility quickly, not the ones that offer the most job security.

Thomas Prommer, Group CEO, We The Flywheel

If you’re hiring 1,000+ grads into an EU or UK organisation right now and your onboarding programme doesn’t include prompt engineering as a core skill, you’re already behind. Not slightly behind. Fundamentally behind.

Here’s the shift I’d frame for any Fortune 1,000 employer: prompt engineering is the new Excel. Twenty years ago, you wouldn’t hire an entry-level analyst who couldn’t build a spreadsheet. Today, you shouldn’t hire anyone who can’t effectively communicate with AI systems to accelerate their output. That’s the baseline. Everything else builds on top of it.

We built Magic Hour to millions of users as a two-person team. David and I don’t have departments. We have AI workflows that do the work of departments. When I think about what made that possible, it wasn’t just technical skill. It was the ability to clearly articulate what we needed from AI tools, iterate fast, and combine that with judgement, creativity, and communication. Those are the exact skills your entry-level hires need.

The ROI case is straightforward. A grad who can use AI tools effectively from day one produces at 2-3x the velocity of one who can’t. Over a 3-year development window, that compounds dramatically. You’re not just saving onboarding time, you’re compressing the entire junior-to-mid career arc. One marketing agency we worked with told me their AI-fluent junior hires were producing client-ready work in weeks, not months.

But here’s what most organisations get wrong: they treat AI skills and soft skills as separate categories. They’re not. The best prompt engineers I’ve seen are actually the best communicators. They know how to break down ambiguous problems, structure their thinking, and articulate intent clearly. That’s a soft skill wrapped in a technical package.

For EU and UK employers specifically, I’d build a hiring programme around three things. First, assess prompt literacy during interviews, not as a gimmick but as a real competency. Second, pair every technical AI training module with collaboration and critical thinking exercises. Third, measure output quality and speed together over the first 90 days to establish your own internal ROI benchmarks.

The organisations that will win the next decade aren’t the ones with the biggest teams. They’re the ones whose smallest teams produce the most. That starts with who you hire and what you consider table stakes on day one.

Runbo Li, CEO, Magic Hour AI

We build software for European clients at Tibicle, and we hire junior developers regularly. One thing I have learned over 12 years of building products is that tools change every six months. The person using them does not change that fast.

When a junior developer joins one of our sprint teams working with a client in the Netherlands or the UK, we do not care which AI tools they know. We care whether they can sit in a standup, explain what they did yesterday, and flag what is blocking them today. That is where most entry-level hires either prove themselves or struggle.

For organisations running large graduate programmes, the skills-first baseline should start with communication and adaptability. Prompt engineering is easy to teach. Knowing how to take feedback, adjust on the fly, and work inside a team rhythm is not.

The ROI over a 3-year programme shows up in retention. Grads who communicate well stick around. The ones who only had technical polish tend to drop off early.

Raj Jagani, CEO, Tibicle LLP

I’ve been hiring and training remote teams for nearly 20 years. Right now I run 15 contractors across six countries. Over 200 people have come through as interns.

Prompt engineering isn’t a skill anymore. It’s literacy. If you’re applying to graduate programmes at Fortune 1,000 companies or government agencies in Europe, everyone in your cohort can use ChatGPT. What matters now is whether you can do what the AI can’t: think systematically, adapt frameworks across unfamiliar contexts, and operate in environments where your home culture isn’t the default.

First reality check: English won’t be enough unless you stay exclusively in the UK. Most Europeans speak two or three languages as standard. US graduates who think fluent English is enough won’t go far. You don’t need fluency in French or German on day one, but you need to be actively learning.

Second shift: ChatGPT is a doer. The new workers need to be thinkers and architects of systems. Can you take a process that worked in one context and redesign it for another? That’s cognitive flexibility, and it separates scaling from repeating.

Third skill: cultural fluency that transfers across borders. This isn’t about knowing facts about other people’s cultures. It’s about receiving other people’s cultural expressions with curiosity and humility, and adapting your communication style in real time. Can you recognise when your American directness is landing as rudeness in a Copenhagen office? Can you adjust your email tone for a Milan client versus a Dublin agency?

An unexpected pattern: people who studied humanities often aren’t credentialed in marketing on paper, but in practice they know how to architect thought processes for clients because of their training. They’ve analysed how arguments are constructed, how narratives shape decisions, and how context changes meaning. That translates directly into cross-border client work.

My advice: stop optimising for credentials and start building proof you can operate outside systems you control. Learn a second language. Work or study where you’re the cultural outsider. Take on a project in an industry you know nothing about and become functional quickly.

The candidates who excel aren’t the ones with polished portfolios, but whose resumes show proof of performance in unfamiliar contexts and can articulate what they learned from not being the expert in the room.

Peter Lewis, Chief Marketing Officer, Strategic Pete