The Skills Nobody Has Yet — And How We’ll Find Them

By Futurist Thomas Frey

How employers will identify, define, and develop the capabilities the future demands — before those skills even have names

A Job Description Written for Someone Who Doesn’t Exist Yet

It’s 2031. A mid-sized logistics company in Columbus, Ohio is trying to hire for a role it’s calling an “AI Operations Interpreter.” The job isn’t about programming. It isn’t about driving. It’s about sitting at the intersection of human judgment and autonomous systems — reading what the machines are doing, translating their outputs for a team of human workers, and flagging the edge cases that no algorithm has been taught to handle.

Six months earlier, this job title didn’t exist. There was no degree program for it. No certification. No LinkedIn skill tag. But the company needed it desperately, so they wrote the description themselves — drawing on a data analyst, a former warehouse supervisor, and a machine learning consultant to figure out what the role actually required.

This is the new normal. And it’s already happening today.

The challenge facing every employer, every educator, and every ambitious professional over the next decade isn’t finding people with the right skills. It’s figuring out what the right skills even are — before the job that requires them becomes urgent.

Why This Problem Is Different From Any We’ve Faced Before

Workforce transitions aren’t new. The industrial revolution wiped out cottage industries and created factory jobs. The computing era eliminated typing pools and created software developers. Every major technological shift scrambles the labor market, and we eventually adapt.

But those transitions played out over decades. A child born into a farming community in 1890 had forty years before the mechanization of agriculture fully restructured rural employment. A typist in 1975 had fifteen years before word processing made her skill obsolete — long enough to reskill.

The AI transition is different because the window is collapsing. Skills that were highly valuable three years ago are already being automated. Skills that will be critically needed in five years haven’t been codified yet. The gap between “this skill matters” and “this skill is obsolete” is shrinking from decades to years — in some fields, to months.

The World Economic Forum’s Future of Jobs Report 2025 surveyed over 1,000 major employers representing 14 million workers and found that 39% of key job skills will change by 2030. That’s nearly four in ten skills that today’s workers rely on, transformed or replaced within five years. The same report identifies analytical thinking, AI literacy, and creative problem-solving as the fastest-rising capabilities — but what’s notable is how few people are being trained in any of them in a systematic way.

So how do we get ahead of this? How do employers identify the skills they’ll need before the need becomes a crisis? And how do workers know what to develop when the target is moving so fast?

 

Signal Reading: How Forward-Looking Organizations Spot Tomorrow’s Skills Today

The companies doing this well aren’t waiting for the labor market to tell them what they need. They’re reading signals — from technology adoption curves, from emerging competitor behavior, from the friction points in their own operations — and working backwards to define the human capabilities those signals imply.

Consider what happened at Amazon. Before drone delivery was operational, Amazon’s workforce planning teams were already modeling what roles would be needed to manage autonomous aerial logistics — not pilots, not warehouse workers in the traditional sense, but people capable of monitoring fleets of autonomous systems, interpreting anomaly reports, and making rapid judgment calls on edge cases. They built internal training programs for roles that had no external hiring market yet, because they knew the external market would take years to catch up.

The same logic applies in healthcare. Radiologists have known for years that AI would handle routine image reading. The forward-thinking hospitals didn’t respond by cutting radiology programs. They asked a different question: what does a radiologist do when the AI flags something unusual and needs a human to make the final call? That question led to a completely new skill profile — less about reading images from scratch, more about supervising and interrogating AI outputs, communicating uncertainty to clinical teams, and making high-stakes decisions under time pressure with incomplete information. Some medical schools are already building this into their curriculum. Most are not.

The Three Lenses Organizations Use to Define Future Skills

From what I’ve observed working with organizations across dozens of industries, the most sophisticated approaches to future skills identification tend to use three distinct lenses — and the organizations that use all three simultaneously are the ones that rarely get caught flat-footed.

The first lens is technology forecasting. You map where the technology in your industry is heading over a three-to-seven year horizon, then ask: what human tasks will this technology automate, what new tasks will it create, and what hybrid roles will emerge at the intersection? This is analytical work, and it requires genuine technical literacy — not deep coding skills, but enough fluency to have an honest conversation about what AI and automation can and cannot do.

The second lens is friction mapping. Every organization has places where work breaks down — where handoffs fail, where decisions stall, where the output of one system doesn’t translate cleanly into the input of the next. These friction points are usually where new skills will be most urgently needed. When a hospital’s AI diagnostic tool flags a result that falls outside its training data, someone has to handle that. When a financial services firm’s algorithmic trading system encounters a market condition it wasn’t built for, a human needs to make a fast call. The friction is the signal.

The third lens is competitive intelligence. If your most innovative competitors are hiring for job titles you’ve never seen before, that’s one of the most reliable leading indicators available. LinkedIn’s labor market data has become one of the most watched signals in workforce planning precisely because emerging job titles cluster in waves — first appearing at a handful of pioneering companies, then spreading across an industry within two to three years. By the time a skill appears in a majority of job postings, you’re already late.

The Skills Taking Shape Right Now

So what does this actually look like in practice? What are the specific skills that are currently moving from “barely mentioned” to “urgently needed” in the labor market?

AI output auditing is one. As organizations deploy large language models in customer service, legal review, medical documentation, and financial reporting, the ability to systematically evaluate AI outputs for accuracy, bias, and appropriateness is becoming a distinct professional skill. It’s not the same as prompt engineering. It’s closer to quality assurance with a domain-specific layer on top — and companies are struggling to find people who can do it well.

Human-machine teaming is another. This is the capacity to work fluidly alongside autonomous systems — knowing when to defer to the machine, when to override it, and how to communicate those decisions to people who don’t share your technical context. It’s part operational skill, part communication skill, and part psychological comfort with ceding control. McKinsey’s research on defining future workforce skills identifies adaptability and comfort with uncertainty as among the fastest-rising needs — and this is precisely why. The people who will thrive are the ones who can hold their judgment loosely enough to update it when the machine sees something they don’t.

Narrative translation is a third emerging capability — and it’s one I find particularly interesting. As AI generates more of the raw data, analysis, and initial drafts across industries, the distinctly human contribution shifts toward interpretation and meaning-making. What does this data actually mean for this specific audience? How do we communicate this risk to people who don’t share our technical vocabulary? How do we make this decision legible to stakeholders with very different frames of reference? These are storytelling skills with professional stakes, and they’re becoming more valuable, not less, in an era of AI-generated content.

 

Refining the Skills: How They Move From Emerging to Essential

Identifying a future skill is only the first step. The harder work is refining it — turning a vague capability into something teachable, assessable, and hireable against.

This refinement process tends to follow a predictable arc. A skill starts as a job task — something specific people are observed doing in high-performing teams. It gets named, usually informally at first, by practitioners inside a company or industry. Early-adopter organizations build internal training for it. Then credentialing bodies, universities, and certification programs formalize it into curriculum. By the time it appears as a standard qualification in job postings, it’s already been through years of informal development.

The organizations winning the talent competition are the ones who enter this arc as early as possible — ideally at the “observed task” stage, before the skill has even been named. Google’s Project Oxygen, which famously studied what made its best managers effective and built training around those behaviors, is a clean example. The skills they identified — clear communication, psychological safety, technical coaching — weren’t invented. They were observed, named, and then systematically developed. The same methodology applies to emerging AI-era skills, just on a faster timeline.

What This Means for the Individual

For anyone navigating their own career through this period, the practical implication is clear: the most valuable thing you can develop isn’t a specific skill. It’s the ability to identify which skills are worth developing, earlier than the people around you.

That means paying attention to where friction exists in your industry. It means reading the job postings at companies two years ahead of yours on the technology adoption curve. It means noticing which conversations in your organization keep hitting the same wall — where the AI output goes, but nobody quite knows what to do with it next. Those walls are where the next round of valuable skills live.

The workers who come out of this transition ahead won’t necessarily be the ones who were best at the old jobs. They’ll be the ones who saw the new jobs coming and started practicing for them before those jobs had titles.

The Bottom Line

The future of skills isn’t a mystery we’re waiting for someone to solve. It’s a signal we can read, if we know where to look. The companies doing this work seriously — mapping technology trajectories, locating friction points, watching competitive hiring behavior — are building talent pipelines for roles that don’t yet exist at scale. The workers paying the same kind of attention are positioning themselves for opportunities that most of their peers haven’t even noticed yet.

The skill that matters most in the years ahead might be the one you’re exercising right now, reading this: the willingness to think seriously about where the world is going, and to start preparing before everyone else catches up.