Synthetic, Not Speculative: A New Way to Prove Creative Work

Opinion piece by Eloise Liley, Chief Strategy Officer, TBWA/Melbourne

Like many in the industry, I’ve spent the past 18 months intensively diving deep into the world of artificial intelligence. How could I not, when LinkedIn feeds and industry headlines are flooded with speculation?

“Will AI replace creatives? Will it automate media buying? Will it erase jobs — or invent new ones?”

But as a pragmatist, I’ve tried to cut through the noise. And what I’ve found is a practical, often overlooked shift happening behind the scenes.

While many are focused on production efficiency or AI’s impact on search, few are talking about innovation in advertising research. Yet this is where AI is delivering some of its most meaningful gains — especially for the C-suite: Chief Marketing Officers, Chief Strategy Officers, and Chief Financial Officers — who increasingly need hard proof that advertising spend delivers measurable value.

In a risk-averse, cost-of-living–conscious environment, Australia remains highly conservative when it comes to advertising. Even as the industry evolves, marketers and media agencies continue to default to television, viewing it as the most bankable and reliable form of media investment.

In many cases, that holds true — TV often delivers the widest reach and the strongest potential for emotional engagement, particularly when addressing brand and audience consideration challenges. But not always.

This continued reliance can become a constraint — especially as agencies and brands explore new formats: new products, branded content, immersive experiences, digital activations, earned activity, and more.

For marketers desperate to innovate, the challenge remains:

How do we prove these alternatives work before committing budget — especially when trying to get TV-agnostic campaigns bought into by the media agencies they collaborate with?

The Rise of Synthetic Data & Virtual Panels

That’s where synthetic data and virtual panels — powered by AI — come in.

Built using real first-party audience data, these research tools simulate real market behaviour. Agencies can now test campaigns and brand experiences in a controlled, data-rich environment. They can generate, refine, and validate new product ideas for brands — something that was previously costly, slow, or impossible pre-launch.

They can forecast:

• Emotional resonance

• Awareness lift

• Likelihood to purchase
  

Without ever going to market.

This isn't theoretical — it’s a growing practice that allows brands to move beyond TV-centric ROI proofing. For finance and strategy leaders, it provides a real framework to assess non-traditional media investment — reducing risk and enabling smarter budgeting decisions.

AI-Powered Creative Testing: Faster, Smarter, Cheaper

While synthetic data opens the door to bold ideas, AI-powered creative testing is slashing the time and cost of validating them.

Australia has long been a market where ad testing is slow and expensive. A single TVC could take weeks and cost tens of thousands to research.

But tools like Kantar’s LINK AI, System1, and local platforms like JunoAI are flipping that script.

Today, brands can test:

• Scripts

• Video edits

• Social ads
  

All in under 48 hours, with predictive results on attention, emotional response, and effectiveness.

Kantar reports that AI can reduce creative testing timelines by over 90%. The financial implications are just as big: tests that used to cost $20,000–30,000 can now be run for a fraction of that — up to 90% cheaper.

That means:

• Campaign ideas tested faster

• More budget redirected to production or media

• Agencies spending less time on research logistics, and more time creating great work
  

For CFOs, it ties marketing spend to outcomes. For CMOs, it boosts confidence in decision-making. For agencies, it clears space for actual creativity.

The Danger of ‘Slop’: When Efficiency Becomes the Enemy

But here’s the caveat:

If we’re not careful, AI’s efficiency gains could accelerate the worst parts of marketing.

We already know the risk: brands over-investing in performance marketing at the expense of long-term brand building. Binet & Field’s 60:40 rule — 60% to brand, 40% to activation — is well known but rarely followed in Australia.

According to Advertising Pays 2023, we spent $17.7 billion on advertising last year — with digital taking over two-thirds. Yet most of that spend went to performance and direct response, not brand equity.

Performance dominates not because it’s most effective, but because it’s most defensible in a quarterly budget meeting.

And now, with AI generating hundreds of ads at scale, the danger is a flood of low-cost, low-impact, forgettable content. I call it “slop.”

AI’s Second Chance — If We Use It Right

But here’s the opportunity:

The same tools that produce low-cost performance ads can also make them better.

AI creative testing can be used not just for TVCs or hero videos, but for:

• Banners

• Six-second cutdowns

• Product-focused content
  

That means brands can raise the creative standard of their performance marketing — making it not just efficient, but engaging. And that improves the overall ratio of investment going toward long-term brand building.

This isn’t a gimmick — it’s a shift in operational models that can deliver real business outcomes.

Final Thought: AI as a Validator, Not a Shortcut

None of this works if we treat AI as a shortcut for making ads rather than a tool for validating them.

In a cautious, cost-constrained market like Australia, AI-powered research offers one of the clearest paths to proving creative value.

After 18 months of deep learning, I believe this is the real opportunity:

To elevate new ideas, validate different formats, and bring creativity back into the business conversation — with proof behind it.

For agencies, it unlocks new ways to innovate.For CMOs, it enables faster, smarter decision-making.For CFOs, it creates a clearer link between investment and outcomes.

Used wisely, AI isn’t just a tool for efficiency — it’s a tool for elevation.