The future has a nasty habit of becoming the past faster than you’d think and, as a result, the “future skills” lists of yesterday tend to age like supermarket sushi. Cast your mind back a few years and everybody was apparently going to become a blockchain developer. Then it was something involving the metaverse. Then prompt engineering arrived and, for approximately eighteen months, appeared destined to become a profession practised by roughly half the population.
The problem with trying to predict the skills in demand five years from now is that we have a tendency to confuse whatever technology is currently making headlines with the abilities that will actually make somebody useful. Those aren’t always the same thing.
As we move towards 2030 and beyond, AI will obviously matter enormously. Recent labour-market research suggests that 39% of the key skills required by employers could change by 2030, with AI, data, cybersecurity and technological literacy among the fastest-rising technical abilities. But sitting right alongside them are creative thinking, resilience, curiosity, lifelong learning, leadership and analytical thinking.
In other words, the future apparently requires us to become more technical and more human at precisely the same time. Helpful.
The UK picture points in much the same direction. Current analysis of the country’s fastest-growing workplace skills puts prompt engineering, large language models and machine-learning operations alongside data storytelling, ethical decision-making, cross-functional collaboration, leadership, negotiation and relationship management. Meanwhile, long-term UK skills research identifies collaboration, communication, creative thinking, information literacy, organising and prioritising, and problem-solving and decision-making as abilities expected to become increasingly important.
That tells us something much more interesting than simply “learn AI”.
It suggests that skill value is being redistributed.
AI is reducing the scarcity of certain forms of execution while increasing the value of the people who can decide what ought to be executed, why, for whom, whether the result is any good and what should happen next. For creatives in particular, that’s quite a substantial distinction.
We’ve already explored why AI literacy is becoming a key leadership skill for agencies, brands and talent, as well as why human creativity could actually become more valuable in the age of AI. The next question is what all of this means for the actual abilities sitting on your CV, portfolio or slightly optimistic LinkedIn profile.
So, rather than pretending anybody can forecast the precise software we’ll all be using in 2031, here’s a Future Skills Index of sorts. It’s not a scientific ranking, but an evidence-based guide to the abilities most likely to gain or lose relative value over the next five years.
At the top of the gaining column sits AI literacy and AI orchestration, as AI moves from specialist novelty towards ordinary professional competence. Judgement and decision-making should rise with it, because cheap generation gives us many more outputs that somebody still has to evaluate. Creative thinking and problem framing become more important for similar reasons: machines are increasingly good at producing answers, while deciding which question is worth answering remains a rather more valuable trick.
Domain expertise should also become more valuable, because specialist knowledge provides the context needed to guide, interrogate and verify AI output. Communication and collaboration are gaining as work becomes more cross-functional and human-AI workflows make explanation and alignment increasingly important. Data and information literacy rise because information itself is abundant while the ability to understand its quality, relevance and meaning is not. Add cybersecurity, governance and risk, plus leadership and strategic thinking, and you’ve got a fairly convincing picture of where the premium is moving.
At the other end, routine clerical and administrative execution looks much more exposed. Generic first-pass production is becoming cheaper as text, images, layouts, summaries and variations become easy to generate. Standalone software operation is unlikely to disappear, but it should become less impressive as a differentiator. The same goes for basic information retrieval. Finding information is getting easier. Figuring out whether that information deserves to be believed is getting harder.
The important word throughout all of that is relative.
“Losing value” doesn’t mean disappearing. Spelling didn’t suddenly become pointless because autocorrect exists. Photography didn’t vanish when cameras became easier to use. Graphic design didn’t end when Canva appeared. What changes is what somebody will pay a premium for.
And that’s where things get interesting.
Why Are Some Skills Becoming More Valuable Than Others?
Gidudu Emmanuel
Skills have economic value partly because they allow somebody to do something that’s difficult, scarce or commercially useful. Technology has always had a habit of changing all three.
Photoshop once made the ability to manipulate an image digitally relatively rare. Eventually, knowing Photoshop became ordinary, and the premium shifted towards knowing what image to make. Something similar happened with web publishing, social media, analytics and increasingly sophisticated production software. AI is simply accelerating the process while spreading it across a much larger collection of tasks simultaneously.
Recent analysis of AI-exposed occupations found that the skills required in the most exposed jobs are changing more than twice as quickly as those in the least-exposed jobs. That might be the most important finding in the entire future-skills conversation, because the real threat isn’t necessarily that your occupation vanishes overnight. It’s that the ingredients required to perform it change beneath you.
A copywriter might remain a copywriter while gradually spending less time producing routine first drafts and more time interrogating briefs, developing concepts, directing AI systems, checking facts, maintaining tone, interviewing subject specialists and defending strategic decisions. A designer could remain a designer while spending less time mechanically creating variations and more time defining visual principles, developing systems, curating outputs, understanding accessibility, directing generated assets and deciding what actually deserves to exist.
The same applies to strategy. AI might consume more desk research and preliminary synthesis while leaving the strategist’s value increasingly concentrated in interpretation, questioning, facilitation and decision-making.
The job title survives. The value moves around inside it.
Research into occupational exposure to generative AI reaches much the same conclusion. Clerical occupations remain particularly exposed, and exposure is increasing across highly digitised professional and technical work too, but most occupations still contain significant tasks requiring human input. Transformation, therefore, looks considerably more likely than wholesale automation across most jobs.
That’s less cinematic than THE ROBOTS ARE COMING FOR YOUR JOB, unfortunately, which is probably why it doesn’t make such an effective thumbnail. It’s also considerably more useful.
The abilities gaining value tend to have a few things in common. They complement technology instead of competing directly with it. They involve context or judgement that can’t easily be reduced to a standard instruction. They create trust and coordination between people. Or they allow somebody to use increasingly powerful systems safely and effectively.
There’s also an interesting divide emerging between work that’s being “professionalised” by AI and work that’s being democratised. In the former, AI strips away lower-level tasks and leaves people concentrating on increasingly expert activities. In the latter, technology makes specialist work easier for non-specialists to perform. The former category has, according to PwC recently shown substantially stronger growth and wage performance.
For anybody wondering how to identify skills for the future of work, that gives us a much better question than “Can AI do this?”
Ask instead: if AI does the easy 60% of this task, what suddenly becomes more important in the remaining 40%?
That’s probably where the value is going.
It’s also why the new skills every creative agency should be building increasingly concern capabilities rather than simply ownership of fashionable software. Knowing how to access AI is becoming ordinary. Knowing where it belongs in a workflow, where it absolutely doesn’t, what constitutes good output and how to manage the risks is much more useful.
Which Skills Will Gain Value Over the Next 5 Years?
Jamie Durrant
Let’s start with the obvious one. AI literacy is going to matter. A lot.
Not because everybody needs to become an AI engineer and certainly not because every creative should spend their evenings memorising prompts like some particularly depressing form of poetry. AI literacy means understanding enough about the technology to use it intelligently.
That includes knowing what different systems can and can’t do, how to structure workflows around them, when human review is essential, how hallucinations and bias can creep in, what information shouldn’t be uploaded, what intellectual property questions need to be asked and when automation is solving a genuine problem rather than merely making a process look futuristic.
Basic AI capability is already moving towards something much closer to general workplace literacy. In early 2026, a major UK training initiative was expanded with the aim of giving 10 million workers practical AI skills by 2030, including foundation-level workplace uses such as drafting, content creation and administrative tasks.
Creativepool has already looked at the AI skills gap facing the creative industries, and I’d expect “AI literacy” eventually to sound about as glamorous as “internet literacy”. Which is to say, not particularly glamorous at all. Just necessary.
Current UK hiring signals are already showing strong growth around prompt engineering, chatbot development, large language models and machine-learning operations, but I suspect tool-specific prompting will ultimately prove less durable than AI orchestration. Today’s magical prompt incantation could very easily become tomorrow’s button.
Understanding how to break a complicated problem into stages, provide useful context, combine systems, verify outputs and introduce human approval is much more likely to survive interface changes. We made a similar distinction in our look at the creative skills agencies, brands and individuals increasingly need: being aware of AI isn’t the same thing as being able to apply it effectively.
Then there’s judgement, which might turn out to be the biggest winner of all.
Generative systems create options astonishingly quickly, which sounds entirely wonderful until you realise that removing one bottleneck tends to create another. If producing twenty concepts takes fifteen minutes rather than two days, somebody still has to decide which of those twenty concepts is actually worth doing.
Which one solves the problem? Which feels distinctive rather than simply competent? Which is legally risky? Which contradicts the strategy? Which will the audience actually understand? Which beautifully phrased statistic has been entirely invented? Which concept seems superficially impressive because the image generator has given everybody cinematic lighting and cheekbones?
The more plentiful content becomes, the more useful discernment becomes. In fact, new tasks appearing in AI-exposed roles have been found to be significantly more likely to depend on human-intensive capabilities such as empathy, judgement and creativity.
That’s a useful counterpoint to the idea that so-called soft skills matter only because humans need something pleasant to do while machines perform the serious work. Judgement is serious work.
As we’ve argued in The Skills That Make Creative Work Distinctive in an AI-Saturated Market, simply getting better at generating things isn’t going to differentiate creatives when everybody has access to increasingly powerful generation tools. What you choose, reject, refine and ultimately stand behind starts to matter much more.
That brings us neatly to creative thinking and problem framing. Creative thinking consistently appears among the capabilities expected to rise in importance over the rest of the decade, while UK research also identifies it as a fundamental employment skill.
For creatives, though, “creative thinking” shouldn’t simply mean generating ideas anymore. Machines can generate ideas. Quite a lot of them. Some are even good.
The higher-value ability is increasingly likely to be framing a problem in a way that produces something interesting in the first place. Anybody can ask an image model for “a futuristic advert for a sustainable trainer brand”. A better creative asks why the trainer brand needs to look futuristic, what sustainability actually means to its audience, which visual clichés the category has already exhausted, what cultural tension the brand might credibly own and whether an advert is even the right answer.
Problem-solving starts upstream.
That’s one reason I remain fairly bullish about taste. Taste is a horribly vague word, admittedly. It sounds like something a creative director says when they dislike a typeface but can’t think of a defensible reason. Proper creative taste, though, is accumulated judgement. It comes from having seen thousands of things, reading outside your discipline, knowing references, understanding what feels fresh and what felt fresh three years ago, recognising when something attractive is actually derivative and noticing cultural signals that aren’t conveniently visible in a dashboard.
That makes curation increasingly interesting as a creative skill. When production becomes abundant, selection becomes labour.
Information literacy and data storytelling are another pair worth watching. We’re hardly suffering from a shortage of information. Quite the opposite. We’re drowning in it, and AI has now supplied several additional fire hoses.
Information literacy is already being identified as an increasingly important employment capability, while UK demand is growing around data-driven decision-making and data storytelling. The useful ability therefore isn’t simply finding information. It’s establishing where it came from, whether it’s reliable, what it really means, what’s missing and how to communicate that to somebody who doesn’t enjoy spending their afternoon staring at seventeen tabs of Excel.
For marketers, strategists, creatives and journalists, this becomes doubly important when generative systems can produce convincing-looking answers faster than anybody can verify them. Research literacy is quietly becoming part of AI literacy.
Cybersecurity, governance and ethical decision-making should gain too. Cyber-risk management, governance, compliance, ethical decision-making and operational resilience are all showing increased demand, while cybersecurity is among the fastest-rising technological skill areas globally.
This doesn’t mean the average art director needs to retrain as a cybersecurity analyst. Nor should they. But creative professionals increasingly handle sensitive client data, unreleased products, customer information, proprietary assets, copyrighted material and AI services whose terms may be considerably less familiar than the software installed on an office computer.
Knowing that “the website let me upload it” isn’t quite the same as knowing you should.
Communication, collaboration and leadership are moving up rather than down as well. Leadership, project management, cross-functional collaboration and cross-cultural communication are all becoming more prominent, while communication and collaboration continue to appear in longer-term projections of future-critical skills.
There’s a fairly simple logic behind that. As individuals gain access to more technical power, organisations need people capable of coordinating that power. A senior creative might now be collaborating with strategists, clients, data teams, developers, AI systems, legal departments and outside specialists on work that once sat relatively neatly inside an agency department.
The ability to explain something complicated without making everybody wish they’d declined the meeting is therefore becoming more valuable, not less.
Finally, learning itself is becoming a skill. Curiosity and lifelong learning repeatedly appear among the abilities expected to grow in importance, while UK labour-market analysis suggests that change has already outpaced earlier projections significantly in some occupational groups.
That’s perhaps the most uncomfortable future skill of all.
You can’t finish it.
Which Skills Will Lose Value Over the Next 5 Years?
Now for the bit everybody secretly came for. What should we stop learning?
Unfortunately, the answer isn’t especially satisfying because very few useful abilities are likely to become completely worthless. What changes is how much scarcity and commercial advantage they provide on their own.
The clearest losers are tasks that are structured, repetitive, digital and relatively easy to verify. Clerical occupations continue to show particularly high exposure to generative AI, while UK research identifies administrative, secretarial, customer-service and certain machine-operator roles among those facing continued pressure.
That doesn’t mean “administration” disappears. It means somebody whose value consists almost entirely of moving predictable information between predictable systems has a more difficult few years ahead than somebody who handles exceptions, relationships, judgement or complicated decisions.
Within the creative industries, I’d put generic first-pass production in a similar category. Basic resizing, routine versioning, background removal, generic social copy, simple summaries, standard presentation formatting, early mood-board generation, transcription, basic research collation, first-pass storyboards and template-led assets aren’t going anywhere.
They’re simply becoming much faster.
And whenever technology makes something significantly faster, the labour market eventually notices. Creativepool’s examination of creative roles facing increased pressure from AI has already looked at what happens when automatable assignments start weakening demand for particular forms of production.
This is especially important for younger creatives because a strange feature of entry-level work has always been that quite a lot of it isn’t terribly exciting. You do junior tasks partly because the business needs them done, but also because doing them teaches you how the business works. AI has the potential to remove some of those apprenticeship tasks before juniors have extracted the learning from them.
There are already signs that expectations are shifting. The most AI-exposed junior roles are becoming far more likely to demand capabilities previously associated with senior employees, including leadership and strategic thinking.
That’s simultaneously encouraging and slightly terrifying.
Apparently the new entry-level requirement is experience.
Excellent.
For agencies and employers, this creates a genuine development problem. If juniors no longer spend several years gradually acquiring judgement while completing simpler production work, businesses will have to teach that judgement deliberately rather than assuming it somehow materialises around somebody’s thirtieth birthday.
Another capability likely to lose relative value is standalone software proficiency. Again, not disappear. Lose value as a differentiator.
Twenty years ago, knowing sophisticated creative software could itself demonstrate relatively rare capability. Today, competence in Photoshop, Illustrator, Premiere, Figma or their equivalents is simply expected for many creative roles. AI accelerates the same pattern.
As interfaces become more conversational and software automates increasing numbers of technical operations, being spectacularly good at remembering which menu contains which command becomes less valuable than understanding why you should use it in the first place.
We made a similar point when looking at the production and post-production skills agencies will need in the coming years. “Prompt engineer” could eventually become one of those wonderfully dated job titles that describes a transitional moment rather than a permanent profession.
I certainly wouldn’t build a five-year career plan around knowing the exact syntax preferred by one current AI model. Learn the principles, learn the workflow and learn the craft underneath it. Interfaces are temporary.
The same caution applies to apparently alarming forecasts around reading, writing and maths. Some projections suggest these foundational skills won’t see the same relative growth in employer demand as AI, cybersecurity, creativity or resilience. That emphatically does not mean writing suddenly stops mattering.
Writing will probably matter enormously.
Being able to produce a grammatically acceptable paragraph on demand is simply becoming less scarce.
There’s quite a difference.
How AI and Automation Are Changing the Value of Skills
Kalai Yung
The strongest evidence that we’re entering a genuine skills transition isn’t simply that AI has started appearing in job adverts. It’s the sheer speed at which job requirements are changing around it.
UK labour-market analysis found an average of 224 newly emerging skills in occupations within the highest quartile of AI exposure, compared with 101 among the least-exposed group, when comparing occupations across the preceding few years.
That isn’t a five-year skills cycle. That’s a skills treadmill.
The same analysis found UK vacancies requiring AI skills rebounded sharply during 2025, while the broader picture shows demand rising simultaneously around AI, data, automation, risk, leadership, collaboration, strategic planning, negotiation and relationship management.
That’s the shape of the transition. AI isn’t simply creating a shiny new bucket labelled “AI skills” and placing it beside all the old ones. It’s changing the relationships between them.
Take writing. AI lowers the cost of producing competent text, so speed of first-draft production loses some scarcity. But briefing becomes more important because vague instructions produce generic material. Editing becomes more important because somebody has to identify the subtly wrong bits. Fact-checking becomes more important because fluent nonsense is still nonsense. Tone judgement becomes more important because the copy needs to sound like the brand rather than the statistical average of the internet. Subject expertise becomes more important because somebody has to recognise when the model has confidently misunderstood the subject.
The same thing happens visually. Generation gets cheaper, so direction gets more valuable. Variation gets cheaper, so selection becomes more valuable. Execution gets faster, so taste becomes more visible.
That’s why the familiar framing of “AI skills versus human skills” is slightly misleading. The strongest workers of the future probably won’t live entirely in either camp. They’ll be the people who combine them.
The designer who understands generative workflows and composition, semiotics, accessibility and brand strategy. The writer who can use an LLM intelligently and interview somebody, interrogate a claim, develop an argument and hear when a sentence doesn’t sound human. The strategist who automates preliminary research and understands consumers deeply enough to recognise when the data is telling an incomplete story.
And the creative director who can scale ideation with AI and still knows when everything on screen is technically polished and spiritually dead.
This is also why AI creativity versus human creativity will probably prove less interesting over time than the enormous grey area of hybrid practices developing between them.
AI might not eliminate many occupations outright. It may simply make the mediocre version of certain skills much cheaper.
That could be every bit as disruptive.
The Skills AI Can’t Easily Replace: Human vs Technical Capabilities
Louis English
So, what skills will be hardest to replace as AI becomes more advanced?
Anybody giving you a permanent list is selling something. There’s no demonstrated collection of completely “AI-proof” intellectual tasks because the capabilities of these systems are changing too quickly for that phrase to mean much.
The more useful question is which skills are currently hardest to commoditise.
The evidence keeps pointing towards capabilities involving judgement, context, interpersonal trust, responsibility and complicated real-world goals. Empathy, judgement and creativity are appearing disproportionately among newly emerging tasks in AI-exposed work, while communication, collaboration, creative thinking, information literacy, planning and decision-making repeatedly feature in UK forecasts. The same pattern is visible internationally, where creative thinking, resilience, leadership, curiosity and analytical thinking are expected to grow alongside technical capability.
That’s a fairly impressive convergence.
For creative professionals, I’d translate all of that research into a few especially durable forms of value.
The first is knowing what problem actually needs solving. AI responds exceptionally well to questions. Working out whether we’re asking the right question remains rather more troublesome.
Clients frequently arrive with proposed solutions disguised as briefs. “We need TikTok.” “We need a rebrand.” “We need to use AI.” “We need something viral.”
The useful person asks why.
Then there’s knowing what good looks like. Generation without standards merely produces more material. A strong creative has references, craft knowledge, taste and enough confidence to reject something that’s perfectly adequate.
Adequacy is going to become extremely cheap.
Distinctiveness isn’t.
That’s also part of the thinking behind Creativepool’s look at the most important skills for building distinctive brands. When increasingly sophisticated tools are available to everybody, clear thinking and human judgement become a much larger part of the differentiation.
Another defensible skill is simply understanding people. Not sentiment scores. People.
What embarrasses them. What reassures them. What they’ll happily say publicly but never privately. Which cultural reference makes somebody feel understood. Why rational behaviour on paper so rarely resembles how humans behave once they’re actually allowed into the equation.
AI can model behavioural patterns impressively. The creative industry will still need people who can enter messy social environments, establish trust and interpret context that was never conveniently entered into a database.
Then there’s taking responsibility, which doesn’t get discussed enough.
Somebody eventually has to say yes. Yes, the campaign can run. Yes, the claim is defensible. Yes, this reflects the brand. Yes, we’re comfortable showing this to millions of people.
AI can provide analysis and recommendations. It can’t absorb organisational accountability in the meaningful human sense. That makes professional judgement particularly important in work involving reputation, ethics, safety or money.
Finally, there’s persuading people. A beautiful strategy that nobody approves isn’t, commercially speaking, a particularly successful strategy. Creative careers involve selling ideas, negotiating resources, handling feedback, reading rooms, resolving disagreements and occasionally explaining for the sixth time why making the logo 300% larger might cause one or two minor problems.
These interpersonal skills don’t sit outside the work. They’re part of getting the work made.
None of which means technical skills are dying. Quite the opposite.
Will technical skills still be valuable in the future? Absolutely.
AI and big data, cybersecurity and broader technological literacy remain among the fastest-growing skill areas, and current UK demand signals include everything from prompt engineering and chatbot development to LLMs, regression analysis and MLOps.
The mistake is assuming technical and human capabilities are competing categories. Increasingly, the premium appears to sit in the combination.
There’s an enormous difference between saying “I know how to use AI” and saying “I understand this market, can identify a meaningful business problem, know where AI could help solve it, can build a useful workflow, know when the output is wrong and can explain my recommendation to a sceptical client.”
One is a tool skill.
The other is professional capability.
That distinction is only going to become more important.
What Makes a Skill Valuable in the Future of Work?
Durgesh Goswami
So, how can you tell whether a skill is likely to remain valuable? I think there’s a reasonably simple test.
Call it the five-year test.
First, does the skill help you decide what matters? Abilities involving prioritisation, problem framing, strategy and judgement sit upstream of generation. Upstream skills tend to retain leverage because they determine where everything else gets pointed.
Second, does the skill make AI more useful rather than forcing you to compete directly with it? Trying to beat a machine at producing fifty routine variations quickly feels like an increasingly questionable career strategy. Knowing how to direct, constrain, integrate and assess those fifty variations is another matter entirely.
Then ask whether the skill is transferable between tools. A workflow principle survives a software update. A keyboard shortcut doesn’t. Understanding how models handle context, how automated systems fit together or how to evaluate uncertainty is much more durable than becoming excessively attached to whichever platform currently enjoys the loudest collection of LinkedIn evangelists.
It’s also worth asking whether your ability requires trust, accountability or genuine human relationships. The greater the consequences of getting something wrong, the more organisations tend to care about who’s actually making the decision. That’s not because humans are infallible. You may have met some. It’s because organisations still require responsibility, negotiation and contextual judgement.
And finally, does the skill rest on real domain expertise?
This might be the safest bet of all.
AI makes generic competence easier to access, which means deep expertise increasingly becomes the thing that guides and checks generic intelligence. A brilliant healthcare creative who understands regulation, patient behaviour and medical communication possesses something significantly more defensible than “good at prompting”. A retail strategist who has spent fifteen years understanding purchasing behaviour has context. A production designer who knows what can physically be built, lit, filmed and delivered within budget has context. A UX designer who understands accessibility and real user behaviour has context.
Domain knowledge lets you recognise outputs that look perfectly plausible to everybody else.
That’s why chasing every fashionable new skill can become counterproductive. Your aim isn’t to collect abilities like Pokémon. It’s to assemble combinations of skills that multiply each other.
A technically strong person with dreadful communication will hit limits. A spectacular communicator with no substantive expertise will eventually be discovered. A talented craftsperson who completely refuses to engage with new technology may find chunks of their workflow steadily commoditised.
The strongest profile is increasingly hybrid.
The emerging roles we’re already seeing across the creative industries point strongly in that direction. As we explored in the best-paid emerging creative roles of 2026, creative technology and AI-oriented product roles increasingly blend technological literacy with communication, commercial thinking, creative judgement and leadership.
And this isn’t only about AI. Environmental stewardship is rising as businesses respond to the green transition. Risk and governance are becoming more important. Demographic change, regulation, cybersecurity, new working patterns and shifting consumer expectations are all colliding at once.
Plus there’ll probably be another platform everybody suddenly insists we need to join.
The safest skill set therefore isn’t one perfectly optimised for a single prediction. It’s one capable of adapting when the prediction turns out to be wrong.
How to Build a Skill Set That Stays Relevant
Loic Jaussaud
The temptation after reading all of this is to sign up for seventeen courses before lunch.
Don’t.
You probably don’t need an enormous collection of new skills. You need a better architecture for the ones you already possess.
For creatives, I’d think about the future skill set as having three overlapping layers.
The first is deep creative or professional craft. You still need to be genuinely good at something. Writing, design, strategy, illustration, animation, direction, photography, development, production, UX, sound, branding. Whatever it is, the arrival of AI doesn’t remove the value of knowing how excellent work is made. In many cases, it makes that understanding more important because the technology lets people without the underlying expertise produce something that looks superficially plausible.
So keep studying the discipline. Look at work. Read. Make things without AI occasionally. Understand the history of what you do and the medium you’re doing it in.
As we argued in Essential Creative Skills for 2035 and Beyond, the future creative isn’t going to survive through technical chops alone, but neither should they abandon craft every time a new tool arrives on a Tuesday.
The second layer is technology leverage, which means becoming genuinely AI literate rather than simply asking ChatGPT to “make this better” and declaring yourself transformed.
Pick a recurring part of your actual workflow. It could be research, concept exploration, asset variation, transcription, data analysis, coding, storyboarding, production planning or personalisation. Then find out what happens when AI is integrated properly.
Where does it save time? Where does quality deteriorate? What information does it need? What still has to be checked by a person? Where are the confidentiality risks? Which parts continue to demand real expert craft?
The goal isn’t to produce an “AI project” purely so you can put the letters AI in your portfolio. It’s to become somebody who understands where the technology creates useful leverage and, just as importantly, where it doesn’t.
Then invest in human leverage. Pick at least one ability that allows the quality of your thinking to travel further: presenting, negotiation, writing, facilitation, leadership, commercial understanding, interviewing, research, stakeholder management or data storytelling.
These skills have the pleasing characteristic of surviving software updates.
The fact that AI-exposed work is increasingly demanding empathy, judgement and creativity, while even junior roles are taking on more strategic and leadership requirements, suggests these aren’t decorative soft skills sitting around the edges of a technical career. They’re moving towards the centre of it.
For somebody trying to act on all this tomorrow morning rather than theoretically preparing for future skills 2030, I’d forget the grand five-year development plan and work in six-month cycles instead.
Pick one technical capability, one human capability and one area of deeper domain expertise. Then build actual evidence that you can use them.
That last part matters. Certificates tell somebody you completed a course. Work tells them you can do the thing.
The creative portfolio of the future will probably need to communicate much more than polished final outcomes. As generative tools make attractive execution easier to produce, portfolios will increasingly need to reveal the problem, your role, the decisions you made, the systems you used, what changed and what value you created.
Anybody can potentially present a beautiful image.
The much more interesting question is: what did you contribute?
Was it your:-
- Thinking?
- Direction?
- Research?
- Craft?
- Prompting?
- Editing?
- Strategy?
- Production?
- System design?
- Decision-making?
That context is becoming important, and it’s also why greater transparency around AI-assisted creative work would be healthy for the industry. Our exploration of whether “Made by Humans” could eventually become a marketing advantage is really another version of the same question. As synthetic production expands, provenance and identifiable human contribution become more interesting, not less.
There’s one other career habit I suspect will become considerably less useful over the next five years: waiting for somebody else to train you.
Long-term UK skills research is already calling for a much stronger culture of lifelong learning precisely because workplace requirements are moving too quickly for the old model of finishing formal education and then coasting on roughly the same knowledge for the next forty years.
Employers obviously have responsibilities here. Particularly if they’re the ones automating half the junior tasks. But individual creatives will also need to become more comfortable with continuous reinvention.
Not constant panic. Not spending every Sunday watching “10 AI TOOLS THAT WILL CHANGE YOUR LIFE” videos created by somebody whose life appears primarily to consist of making videos about AI tools.
Just deliberate learning.
And there’s something reassuring buried underneath all of this. The evidence doesn’t actually suggest that everybody needs to throw away what they’re good at and retrain as a machine-learning engineer. Quite the reverse.
Across the research, a surprisingly consistent picture emerges: the skills in demand are becoming combinations of technological capability, human judgement, adaptability, communication and expertise.
Technical skills will matter. Soft skills will matter. Creative skills will matter. Domain knowledge will matter.
What’s changing is how they fit together.
The person most exposed to disruption isn’t necessarily the least technical person in the room. It’s the person whose entire value can be described as a repeatable task.
The person in a much stronger position is somebody who can understand a messy problem, bring specialist knowledge to it, use powerful technology intelligently, recognise what good looks like, work effectively with other people and ultimately take responsibility for the result.
Which, admittedly, is quite a lot.
But it’s still considerably more reassuring than being told to spend the next five years memorising prompts.
Because the most valuable skills for the future of work probably won’t be the ones permanently attached to a particular piece of software.
They’ll be the ones that help you decide what technology should do.
