Skills for Tomorrow

50% of the roles we're recruiting for today won't exist in 5 years. The other 50% will be completely transformed.

Yes, these are arbitrary numbers, but given that it’s impossible to scientifically estimate the impact of AI in the future, we feel this is a fair - if broad - guess of what’s to come.

I know what you’re thinking: hype. It’s the fear/uncertainty/doubt playbook that’s being peddled by all of the AI labs. And normally, we would share your cynicism.

But from what we’re seeing, it’s not hype. It’s the reality we’re witnessing firsthand in tech recruitment as we hit mid-2025.

There are plenty of companies chasing "Senior Python Engineers with 8+ years in legacy frameworks" while their rivals have started building AI agents that handle coding tasks autonomously. There are firms scouring the market for "QA Testers" when AI can debug and test software in real-time (albeit, with training and plenty of supervision - for now).

There are many firms still filling yesterday's positions while tomorrow's workflows are being redefined by AI at breakneck speed.

 Companies seem fixated on certifications that time-out before onboarding even wraps. Just last week, a client mandated "expert-level proficiency in traditional SQL databases" for a data innovation role, ignoring that AI-driven tools and vector databases are now standard for querying massive datasets in seconds.

The real skills gap isn’t a dearth of folks who can spin up Docker containers or wrangle APIs. We're critically short on talent that can co-pilot with AI: mastering multimodal prompting, integrating Grok-like models into workflows, and strategising as AI-human hybrids for complex challenges.

Now you’re probably thinking “yes, but hiring AI-native is fine for start-ups and those without legacy estates. What about that?”. And it’s true - the whole AI thing is being delivered with a large dollop of false prophecy…tomorrow will be better, but don’t worry about how we’ll get from here to there. Just trust us.

 In the meantime though, old-school HR is keyword-matching resumes from 2020. But the MVPs of late 2025 are the hybrids who fuse human intuition with AI augmentation; people discovered through demonstrating their practical application of technology (probably in a hobbyist setting, at least in part - because these people are real enthusiasts, and not just building stuff to build their career).

So what do companies really need now?

Systems Architects are needed more than task executors - people who map how AI ecosystems mesh with business ops, designing resilient flows that blend human ingenuity with machine precision.

And we’ve all heard about the much-vaunted “prompt engineering” role; those who don't just query AI but engineer problem frames for optimal outputs.

 You’ll also need some visionaries who safeguard the irreplaceable: empathy-driven client interactions, high-stakes ethics calls, clear and unambiguous business analysis, innovative leaps in uncertainty…all domains where AI enhances but never replaces, and hopefully never will.

Instead of seeking candidates with X years’ experience in Y cloud platform, it’s more crucial to find people who can quickly pick-up the basics of any stack/platform and orchestrate AI to shoulder the grunt work. It’s better to prioritise curiosity over tenure - seeking those buzzing about breakthroughs like the latest models, not those who've plateaued on outdated tech.

If you're building teams today, it’s probably best to ditch the 2020-era JDs. Envision 2027's landscape - AI ubiquity, agentic systems, quantum edges. Shift from "Fit for now?" to “Adaptable as AI reshapes this?"

And if you're navigating the job market, beyond mastering the core technology trends, you could learn to amplify yourself via AI. What can you do faster, better with it, than the next person without?

Firms clinging to legacy hiring will end up with teams ill-equipped for AI-augmented futures.

Those nailing it will surge ahead with lean, supercharged squads, outpacing the manual laggards of today.

Hire for evolution, not extinction. Or get the experts in to help (that’s us, by the way).