AI has now reached that slightly awkward stage where everyone has officially stopped pretending it’s a novelty, but nobody has quite agreed what to actually do with it.
According to the UK government’s June 2026 AI adoption plan, 51% of creative businesses now report using AI, compared with 33% of businesses across the wider economy. In design and designer fashion, adoption sits at 53%. In film, TV, video and photography, it reaches 44%.
So no, this is not a fringe experiment being carried out by three overexcited futurists and a brand consultant with a headset mic. It’s infrastructure now. Or at least, it’s trying very hard to become infrastructure.
That matters because the old question was always a bit too blunt. “Will AI replace creativity?” is the kind of question that sounds dramatic on a panel but doesn’t actually get anyone very far. The more useful question is what happens to the creative workflow when ideation, versioning, localisation, resizing, tagging, transcription, production support and adaptation all become much faster.
Microsoft’s 2025 Work Trend Index describes the emerging model as “AI-operated but human-led”, while WPP now frames its own platform strategy around end-to-end workflows in which “AI and human creativity work as one”. In other words, hybrid human-AI systems are not a distant future waiting politely in the wings. They’re already turning up to the meeting, connecting to the Wi-Fi and asking who owns the deck.
But speed is only the easy part of the story. The harder question is control.
As Nadim Chartouni, Executive Producer and Business & AI Strategist at Tonic DNA, puts it:
“Everyone's talking about AI making content cheaper and faster. Nobody's talking about what happens to the feedback loop.
In traditional animation production, for example, client control is baked into the process, storyboard approvals, animatic reviews, character design sign-offs. Every stage is a checkpoint where a client can say "more like this, less like that." That control costs money. It's built into the budget.
So, here's the real question as AI pipelines compress timelines: if budgets shrink in exchange for speed, do clients actually want less control? Or do they want the same control, just exercised differently?
We don't think the answer is "less oversight." We think it's "oversight at different points."
When generation is fast and cheap, the temptation is to skip straight to a finished-looking output and ask for a thumbs up or down. But that's actually less control, not more, you're reacting to a result instead of shaping a direction.
The real opportunity with AI isn't removing checkpoints, it's moving them earlier: more iteration on look, tone, and narrative intent before anything gets "finished," because iteration itself is now nearly free.”
That is probably the neatest summary of where the creative industries are heading. Not towards a world with no human judgement, but towards one where judgement must happen earlier, more often and with far more intent.
The Traditional Creative Workflow Is Starting to Look Very Tired
Shafaq Pervaiz
For decades, the creative workflow was basically linear. Brief. Concept. Production. Review. Delivery. There were detours, obviously, because creative work has never moved in a perfectly straight line unless somebody has aggressively removed all the interesting bits. But the structure broadly held.
Each stage was a gate. Each gate cost time. Each round of feedback carried production consequences. In animation, design, advertising, film, content and brand work, that structure was not just creative process. It was commercial architecture.
Now AI is making that architecture wobble.
Adobe’s Firefly relaunch in April 2025 positioned the platform as an all-in-one environment for ideation, creation and production. By June 2026, Adobe had expanded its “creative agent” across Photoshop, Premiere, Illustrator, InDesign and Frame.io, describing it as connective tissue across different stages of creative work. Figma is moving in a similar direction, with AI tools designed to take teams “from first spark to shipped product”.
That phrase is doing a lot of work. “First spark to shipped product” is not a tool feature. It’s a worldview. It imagines the creative process less as a chain of departments and more as one continuous system.
Linear workflows are giving way to agile, iterative processes and hybrid teams in which humans set direction while AI agents take on more execution. Put those signals together and the old pipeline starts to look rather antique. The creative workflow is becoming less like a production line and more like a set of loops: between brief and prototype, draft and approval, data and judgement, scale and taste.
That sounds efficient. It may even be efficient. But it also means creative teams must rethink where decisions belong.
If production becomes cheaper, the temptation is to rush towards something that looks finished and then invite people to approve it. That feels fast. It also risks being a terrible way to work. Because once something looks finished, people respond to it as a finished thing. They react. They nitpick. They ask for the blue to be warmer and the button to feel more “confident”. Everyone starts renovating the wallpaper while the foundations quietly shift underneath.
The better opportunity is to move the meaningful conversations earlier. What should this feel like? What is the tone? What should the audience understand in the first three seconds? What are we definitely not making? What kind of world does this brand belong in? What does “premium” mean here, and does anyone in the room agree?
AI can make those conversations more visual, more iterative and more affordable. But it can’t make them unnecessary.
The Real Shift Is From Production Control to Directional Control
Phenomenon Studio
This is where the AI debate has often been painfully thin. Too much of it has been obsessed with output. Who made the image? Who wrote the line? How fast was the video produced? How many variants can the team generate by Friday?
Those questions matter, but they are not the whole machine. The creative industries do not run on outputs alone. They run on taste, alignment, trust, approvals, rights, reputation and the delicate shared fiction that everyone knows what “make it more human” means.
In the old model, control was visible because it was staged. Storyboard approvals. Animatic reviews. Character design sign-offs. Script revisions. Rough cuts. Fine cuts. Final cuts. These were creative moments, but they were also commercial protections. They gave the client confidence that the work was moving in the right direction before too much money had been burned.
AI threatens to scramble that logic.
If a team can generate ten polished-looking routes in an afternoon, it may look as though control has increased. More options. More speed. More surface area. Lovely. But more output is not automatically more control. Sometimes it’s just more stuff to be confused by.
The better hybrid workflow gives clients and creative teams more influence over the direction before the work hardens into fake finality. It uses AI to explore mood, tone, narrative, pacing, visual language and adaptation while those things are still cheap to change.
That’s the difference between using AI as a vending machine and using it as a thinking environment.
Feedback cycles remain one of the most time-consuming parts of the content supply chain. Research into human-AI co-creativity also suggests that systems with higher user control tend to create greater trust and ownership.
That should surprise nobody who has ever sat in a creative review. People are far more likely to trust a process they can shape. They are far more likely to back work when they understand how it got there. And they are far less likely to panic at the end if they were part of the decision-making at the beginning.
What Hybrid Human-AI Creative Workflows Actually Look Like
Juice
The strongest hybrid workflows do not begin with a prompt. They begin with direction.
That may sound boring, but it’s the whole game. AI can produce options at speed, but somebody still has to know what the options are for. Good creative teams will need sharper briefs, stronger references, clearer constraints and more confident taste at the start of the process.
That feels about right. Prompting is useful, but prompting without judgement is just ordering from a menu in a language one barely speaks.
In practice, AI is already proving valuable in the parts of the workflow where volume, variation and speed matter. Adobe’s Firefly Boards support mood boarding, brainstorming and rapid exploration. Firefly’s broader platform spans image, video, audio and vector generation. Figma’s AI, meanwhile, allows teams to move from prompts or existing designs into design and code.
That makes AI especially useful for early concept expansion, visual exploration, repetitive production tasks, structured adaptation, localisation and multi-format deployment.
The production layer is where the business case becomes more obvious. Adobe’s enterprise tools now emphasise governed workflows that combine models, creative actions, reviews and delivery steps across a full production lifecycle. The platform is aimed at campaign variants, localisation, e-commerce imagery and large-volume output, with logs, histories and centralised oversight.
That is not glamorous in the Cannes Lions sense. Nobody is going to make a stirring montage about asset tagging. But in real creative businesses, this is where a lot of time goes to die.
This is also where the most sensible AI argument naturally lives. Not “the machine will make the idea”, but “the machine can reduce the drag between an approved idea and all the places it needs to show up”.
Humans Still Decide What Good Means
Publicis Sapient
For all the noise, AI’s strengths are relatively easy to describe. It is good at pattern extraction, structured transformation, high-volume variation and repetitive execution. Microsoft found that nearly half of 365 Copilot conversations supported cognitive work such as analysing information, solving problems, evaluating and thinking creatively. Adobe’s assistants follow a similar logic, helping organise, rename, classify, generate starting points and manage production tasks. That is useful. Very useful. It is also not the same as taste.
In other words, AI can help make more. It cannot reliably decide what deserves to exist.
That distinction is becoming more important, not less. Creative quality is not just novelty. It is fit. It is timing. It is tone. It is emotional calibration. It is knowing when the obvious line is actually the right one and when the clever line is quietly killing the idea. It is understanding brand voice, cultural context, legal risk, audience sensitivity and the unspoken difference between bold and embarrassing.
Which is inconvenient, perhaps, for anyone hoping to replace creative judgement with a subscription plan. But it is very good news for people who can actually think.
The Risk Is Not Automation. It’s Passive Creativity
Matt Chansky
There is another issue here, and it is less discussed because it does not fit neatly into a productivity deck.
Research has found that human-GenAI collaboration improves immediate task performance but can also reduce intrinsic motivation and increase boredom when people return to solo work. That finding should make creative businesses pause. Not because AI is inherently bad for creativity, but because badly designed workflows can make creative people feel less like makers and more like quality-control staff for a machine with confidence issues.
If AI turns writers, designers, animators, editors and strategists into passive approvers of machine output, the work may get faster and worse at the same time. People may produce more while caring less. They may become fluent in tweaking and rusty at originating. They may spend their days choosing between variants rather than developing a point of view.
That is not a future worth sprinting towards.
The better model uses AI to preserve agency. It gives creative professionals more room to explore, compare, direct and refine. It lets them test options without pretending all options are equal. It allows clients to see possibilities earlier without mistaking early polish for final quality.
This is why the phrase “AI does the boring bits” is only half true. Some repetitive tasks absolutely should be automated. But creative work is not cleanly divided into boring execution and noble inspiration. Exploration, comparison, curation and revision can be laborious, but they are also where much of the craft lives.
When AI enters those spaces, it does not just save time. It redistributes attention.
The question is where that attention goes next.
How Agencies Are Adapting
Nalla Design
The smartest agencies are not treating AI as a decorative extra. They are rebuilding around it.
WPP says its platform connects the entire marketing workflow, from strategy to creative, media and production, through AI agents guided by human talent. Its 2025 agency report argues that agility, real-time collaboration and integrated workflows are replacing siloed structures and fixed deadlines.
That is a meaningful shift. The agency value proposition is moving away from pure labour-hours and towards orchestration. Who can combine tools, people, systems, approvals, rights, data and judgement into a faster, more adaptive service without turning the work into generic paste?
This is why so much AI adoption clusters around production friction. Versioning. Localisation. Asset management. Review. Searchability. Design-to-code continuity. Campaign variants. E-commerce imagery. These are not side quests. They are the machinery of modern creative delivery.
But the client relationship needs to change too.
If AI compresses timelines and budgets, clients are unlikely to say, “Wonderful, in that case we’ll accept less oversight.” That is not generally how clients behave, because clients are not fictional woodland spirits from a productivity manifesto. They still want control. They still want confidence. They still want to know the work won’t humiliate them in public or vanish into the beige middle of the internet.
The sensible shift is towards earlier, better, more visible oversight. More alignment before production. More iteration before commitment. More clarity around what is being approved and why. In this landscape, there’s a real need for centralised visibility, simplified approvals and connected collaboration rather than scattered comments across half a dozen tools.
Hybrid human-AI systems are therefore becoming the new standard not because everyone has solved them, but because the incentives are increasingly unavoidable. The UK government’s adoption plan explicitly supports an “augmentation first” approach, helping creative workers spend more time on the creative aspects of their jobs.
Still, the shift is uneven, and it should be. The government’s sector plan notes that many firms remain cautious where client-facing work intersects with intellectual property and ethics. The ILO has warned that GenAI in media and culture raises serious questions around job exposure, employment conditions, compensation and creative control.
That is not resistance to progress. That is the industry noticing that speed without trust can become a very expensive shortcut.
The New Skills Are Less About Prompting and More About Judgement
Laundry
If AI changes the role of designers, writers, marketers and creative directors, it does so less by replacing expertise than by moving expertise around.
Prompting matters, obviously. But “prompt engineering” is a strangely thin way to describe what creative people actually need now. The more valuable skill set is broader: defining intent, structuring inputs, evaluating outputs, diagnosing drift, understanding model limits, protecting rights and building review systems that keep quality high under pressure.
AI-exposed jobs are changing more than twice as fast as less exposed roles, and skills such as judgement and leadership are becoming more critical and more rewarded. That feels like the real brief. Not simply “learn the tools” but learn how to remain useful when the tools become normal.
The emerging skill is not using AI. It is using AI inside a governed creative system.
That means knowing when to trust the tool and when to ignore it. It means knowing how to document decisions, maintain brand consistency, spot legal risk, protect source material and prevent fast work from becoming shallow work. It means being able to explain to a client not just what was made, but how it was made and where human judgement shaped it.
The UK government’s adoption plan calls for trusted standards, business-friendly tools, leadership development and practical AI capability programmes for workers and freelancers. It also highlights the need for hybrid creative-technical roles that bridge creative practice and technology.
That may be one of the most important ideas in the whole debate. The most useful person in the room may not be the loudest AI evangelist or the most committed purist. It may be the person who can translate creative ambition into a workflow that scales without flattening the work.
Freelancers and Smaller Studios Need Protection, Not Just Tools
Torpedo
This point matters because the creative industries are not made only of giant agencies and enterprise platforms. They are also made of freelancers, small studios, production specialists, illustrators, editors, designers, animators, photographers, copywriters, composers and all the people who form the flexible, slightly precarious backbone of the sector.
Freelancers across the creative industries (myself included) are arguably on the front line of AI disruption and often lack the protections of salaried workers. That is the bit the AI hype train often skips over. Tools arrive quickly. Transition arrives slowly. Contracts lag. Rates get squeezed. Expectations rise. Clients ask for more versions, faster, for less. The language of empowerment starts to feel suspiciously like a discount request with a keynote attached.
If creative businesses want to adopt AI without hollowing out creative quality, they cannot only invest in software. They must invest in people, governance, fair transition and clear rules around rights and attribution.
Otherwise, the sector risks gaining efficiency while eroding the human capability it still needs to make the work worth paying for.
The Future of Creative Workflows Is Hybrid, But Not Settled
ROCANI
The future of creative workflows is unlikely to be fully automated. It is also unlikely to be the pre-AI status quo with a chatbot taped to the side.
The more plausible model is augmentation-first: AI expands options, accelerates iteration and handles structured production work, while humans retain authorship over direction, selection, ethics, taste and accountability.
That is a strong model when designed well. It is also full of unresolved problems.
The UK government’s March 2026 copyright report acknowledges that frontier AI models require billions of inputs, many of them copyrighted works. It also notes growing concern around realistic impersonation and digital replicas. The House of Lords Communications and Digital Committee has been even blunter, warning that speed is not a substitute for human creativity and raising concerns around opaque, unlicensed training regimes.
These are not abstract legal wrinkles. They will shape what clients allow, what agencies promise, what creators accept and what hybrid workflows become commercially viable.
So yes, hybrid human-AI workflows are becoming the new standard. But “standard” does not mean settled. The standards themselves are still being negotiated: copyright, provenance, disclosure, client consent, compensation, quality assurance and worker transition.
The winner will not simply be the fastest pipeline. It will be the one that combines speed with trust, control and discernment.
That brings everything back to Chartouni’s central point. Someone still has to decide what’s good.
AI changes when that question gets asked. It changes how many times a team can ask it. It changes how cheaply different answers can be explored. But it does not answer the question on behalf of the creative industries.
The next creative workflow, then, is not post-human. It is more deliberately human. More dependent on framing, taste, editorial courage, client dialogue and the ability to move feedback to the point where it can actually shape the work.
AI may make content cheaper and faster. Fine. But cheaper and faster is not a creative philosophy. It is a production condition.
The real opportunity is not to remove people from the loop. It is to build a better loop: one where human judgement shows up earlier, matters more clearly and has more room to do the thing creative judgement has always done best.
Decide what is worth making in the first place.