AI did not replace the creative process. It made a new one possible.

For the past year, most conversations about generative AI in creative production have centred on what the technology can do.

Can it generate a film? Can it create product shots? Can it produce something that looks cinematic? Can it reduce the cost of production?

The answer, increasingly, is yes.

But that is no longer the most interesting question.

The more important question is whether AI can produce work that is good enough to carry the weight of a serious brand.

Because for enterprise organisations, near enough is not always good enough. A concept might be 90% there. A shot might feel impressive at first glance. A visual might look cinematic until the logo warps, the wheels spin backwards, the lighting feels off, or a niche technical detail is guessed incorrectly.

That was the real lesson from a recent project PLAYE delivered for Fortescue.

The work used generative AI to create an ambitious brand video that would have previously required either a major live-action shoot, a complex CGI production, or would simply never have made it past the idea stage.

But the story is not that AI made the video.

The story is that AI made the idea possible, and human creative craft made it work.

‍

The brief: an ambitious idea with real-world constraints

Fortescue came to the project with a clear creative direction and a storyboard already developed with the help of AI. That gave the team a strong starting point.

PLAYE Creative Partner Keir Crighton assembled an expert team led by Senior Creative Director Paul Mansfield. This team, including an editor and sound engineer, transformed the initial storyboard into a high-impact cinematic campaign by refining transitions and visual style while balancing creative ambition with budget and time constraints.

The concept included highly stylised environments, fast-moving vehicle sequences, mining infrastructure, Formula E-related requirements and specific brand elements that needed to appear correctly on screen.

This is where generative AI becomes both powerful and imperfect.

If the brief asks for something imaginative, surreal or invented, AI can be extraordinary. It can create worlds that do not exist. It can generate atmosphere, scale and visual drama quickly.

But when the brief asks for something specific, niche or technically accurate, the challenge becomes much harder.

A generic race car is one thing. A particular car with specific logos in specific places is another. A generic plug is one thing. A specialised mining charging plug is another.

AI is only as strong as the data and context it can draw from. The more niche the subject matter, the more likely the model is to guess.

And when it guesses, it can get things wrong.

Why the storyboard mattered more than ever

One of the biggest learnings from the project was that AI does not reduce the need for planning. In many ways, it makes planning more important.

The storyboard became the creative anchor for the entire production. It helped define the sequence, style, camera movement, rhythm and visual logic of the piece before the team moved into generation.

As Paul described it, changing the storyboard early is like changing an architect’s drawing before construction begins. It is far easier to adjust the plan before the build starts than to tear work down halfway through.

That matters because AI production still carries real production costs. Not only in dollars, but in time, credits, iterations, creative energy and quality control.

For this project, around 1,200 images were generated. That volume was not a sign of inefficiency. It was part of the quality control process.

The value came from generating enough options to find the strongest frames, then refining, rejecting and rebuilding until the work felt cohesive.

AI gave the team range.

The creative process gave the team judgement.

The difference between AI output and finished work

One of the most useful ideas to come out of the project was that AI is not really replacing the creative process. It is changing where the creative process happens.

In a traditional shoot, a director, cinematographer and production team make hundreds of decisions before and during production.

What is the subject? Where is the camera? What is the time of day? What is the light doing? Is the scene handheld or smooth? Does the car feel fast? Does it hug the track? Does the body move like it has real suspension? Does the environment feel believable?

Those decisions still exist in AI production.

They just happen through prompts, references, iteration and post-production.

For one moving car shot, for example, it was not enough to generate a car on a track. The vehicle needed to feel physically connected to the road. It needed to bounce in the right way. The wheels needed to behave properly. The lighting needed to make the car pop without making the scene feel artificial.

That is not a shortcut around craft.

That is craft, expressed through a new production medium.

The problem with “AI slop”

There is a lot of AI-generated content in the market right now.

Some of it is impressive. A lot of it is not.

The risk for brands is not that audiences will always know something was made with AI. The risk is that the work feels careless, cheap or disposable.

That is where the word “slop” becomes useful.

The issue is not whether AI was involved. The issue is whether the work has been made with care, judgement and enough attention to detail.

AI slop happens when outputs are accepted too quickly. When errors are missed. When logos morph. When lighting feels wrong. When the music, edit or sound design fails to elevate the piece. When nobody has taken responsibility for the final result.

For brands like Fortescue, that distinction matters.

Enterprise brands cannot simply publish something because it looks impressive at first glance. The work still needs to be checked, shaped and refined. It needs guardrails. It needs creative direction. It needs someone who can spot the difference between “good enough for a generation” and “good enough for the brand”.

That is where the human team becomes essential.

AI can get you somewhere. Craft gets you over the line.

One of the more interesting tensions in the project was how far AI actually gets you.

In one sense, AI made the impossible possible. Without it, producing this kind of film may have required a major physical shoot, expensive CGI, specialist crew, location access, vehicle logistics and a much larger budget.

In that context, AI closed a huge gap.

But inside the AI workflow itself, the final stretch was still demanding. The team still needed to close the gap between generated output and polished production.

That meant refining shots, rebuilding frames, controlling movement, managing visual consistency, correcting errors and adding the final layers of sound design and editing.

The result was not “AI did the work”.

The result was a new kind of production model, sitting somewhere between traditional production and CGI.

It gave the team the ability to create something ambitious without needing to physically shoot every element. But it still required the mindset of a filmmaker, the eye of a creative director and the discipline of a production team.

What this means for creative teams

The Fortescue project points to a more interesting future for AI in creative production.

It is not simply about doing the same work cheaper.

It is about unlocking work that previously may not have happened at all.

Ideas that were once too expensive, too complex or too hard to justify can now be explored in new ways. A brand can create cinematic worlds without a full-scale shoot. A social concept can be elevated beyond what the budget would have traditionally allowed. A team can test, generate and refine at a speed that changes what is possible.

That does not mean creative roles disappear.

It means they change.

The value moves towards people who can direct the tools, control the output, understand brand risk, spot the errors, apply taste and know when the work is genuinely ready.

Anyone can generate something.

Not everyone can make it good.

The takeaway

The Fortescue project was not interesting because AI was used.

That is becoming less remarkable by the month.

It was interesting because it showed what happens when AI is treated as a production capability, not a gimmick.

The technology helped unlock an idea that would have been difficult, expensive or impractical to produce through traditional methods alone. But the quality came from the people around the tools.

The storyboard mattered.

The prompting mattered.

The references mattered.

The quality control mattered.

The edit, music and sound design mattered.

The creative judgement mattered.

That is the real opportunity for brands.

AI can generate the raw material. But the difference between disposable AI content and brand-worthy creative work still comes down to craft.

And in a market filling quickly with AI slop, that difference is only going to matter more.

‍