Tom Morton, head of activation at Door4, considers how predictive AI is transforming digital marketing – and whether creativity might suffer as a result

Experimentation has always been central to performance marketing.

At Door4, we constantly test and learn, from fine-tuning PPC campaigns to refining landing pages for conversion.

Sponsored Video

AGA’s new colour, La Cornue bling & Novy’s hidden hob | Clerkenwell 2025

The traditional method has always been hands-on: try something, measure the impact, roll it out if it works.

But that’s changing fast. The rise of predictive AI is replacing some of this real-world trial and error with simulated outcomes.

It’s promising faster results and fewer wasted clicks, but it also begs the question: are we losing the creative spark in the process?

Human touch

PPC, SEO and CRO have always relied on a mix of data and gut instinct. We test ad copy and audiences.

We tweak meta tags and layouts. Sometimes the wins are logical, and sometimes they’re a complete surprise.

And that’s the point. The unpredictability of real experimentation is what drives so many breakthroughs.

A quirky PPC headline that beats the safe option. A layout that turns more browsers into buyers than anyone expected. These moments of surprise often push a campaign from good to great.

With machine learning, AI tools can now simulate campaign results before anything even goes live.

In PPC, they can model how different ad creatives might perform.

In SEO, they can predict the impact of new content on rankings.

In CRO, they can estimate which layout or UX changes will move the needle.

That means no more waiting weeks for data, and no more budget wasted on duds. On paper, that’s a dream come true.

But if we rely too heavily on predictive models, we risk filtering out the ideas that don’t fit the algorithm.

A bathroom brand playing with tone or humour might never test a quirky idea if the model predicts it won’t perform.

A kitchen retailer trialling a bold layout or unexpected UX tweak might never discover the upside if AI says “stick with the norm.”

And while AI can simulate behavioural patterns based on historical data, it can’t predict cultural shifts or emotional responses in real time.

Sometimes audiences defy expectations – and that’s when marketers really learn something.

Balance matters

There’s a strong case for using predictive AI to reduce waste and sharpen decisions.

For brands with smaller budgets – and less time to experiment – AI offers a smart way to optimise without the trial-and-error cost.

But that doesn’t mean traditional testing is dead.

If anything, the smartest brands will blend predictive tools with instinctive, creative testing.

They’ll use AI to avoid obvious failures and explore more radical ideas in parallel.

Predictive models show us what’s likely to work. It’s still up to us to decide what’s worth doing.

As predictive AI becomes more accessible, marketers will spend less time tweaking campaigns and more time interpreting models and spotting opportunities.

The job isn’t to blindly follow the data – it’s to know when to challenge it.

Creativity won’t disappear. But it will need to work harder.

Because in a world where AI can build the best-performing version of “average,” the biggest wins will come from marketers who know when to go off-script.