Creative Production

Breaking Gen AI’s efficiency trap: Reuniting creative and media

Pencil Performs

The technology exists. Platforms like Pencil have already processed billions in media spend, generating millions of assets that are predicted for performance before launch, automatically formatted for every platform, and continuously optimized based on real results.

Jack Smyth is Country Lead for Jellyfish + Brandtech Group in Australia

Here’s the uncomfortable truth about Gen AI in marketing: the technology works as promised.

It’s generating 10x more assets at 10x lower cost, but for some brands that barely moves the needle.

Brands are caught in what I call the Silicon Ceiling—they’ve deployed AI inside their organizations to make creative production faster and cheaper, but they’ve forgotten to connect it to the thing that actually delivers business growth: media.

They’re saving production dollars but never reallocating those savings into working media. They’re generating more creative without any mechanism to know what works. They’re stuck optimizing for efficiency while their competitors optimize for effectiveness.

Faster, cheaper creative that doesn’t work is just faster, cheaper failure.

The Broken Handoff That AI Finally Fixes

For decades, the advertising industry operated with a fatal disconnect. Our original sin was splitting creative and media. Creative agencies built beautiful, award-winning work based on intuition and “brand truth.” Media agencies took that work and tried to make it perform, often discovering too late that the beautiful creative didn’t convert.

By the time performance data revealed what worked, it was too late—the creative was locked, the budget was spent, and everyone moved on to repeat the same broken process.

Gen AI doesn’t just make this workflow faster. It obliterates the handoff entirely.

The real revolution isn’t speed or cost. It’s that creative and media can finally speak the same language: performance. AI makes it possible to generate creative variations informed by actual media data, predict performance before launch, and continuously optimize based on real results. Media insights don’t just optimize targeting—they optimize the creative itself.

This is the shift from the Silicon Ceiling to the performance breakthrough: reuniting creative and media into a single, intelligent system.

What Intelligence Actually Looks Like

Progressive marketers aren’t asking their Gen AI platforms to be cheaper Photoshop. They’re asking them to be smarter strategists.

Consider what happens when your creative platform is built on over $2 billion in actual media spend across thousands of brands: it knows that product-forward imagery outperforms lifestyle shots by 40% in your category during Q4. It knows certain color palettes drive 2x better engagement with your target demographic. It knows videos under 8 seconds significantly outperform 15-second spots on specific platforms.

That’s not science fiction – it’s how thousands of brands are using our platform Pencil.

This isn’t replacing creativity—it’s arming creatives with intelligence they’ve never had before. The best creative directors have always had an intuitive sense for what works. Now they have the data to prove it, scale it, and continuously improve it.

But here’s the critical part: this only works when your Gen AI platform connects directly to media channels. If you’re generating creative in isolation, running it manually through approval workflows, and uploading it separately to Meta, TikTok, Google, and Amazon, you’re still stuck at the Silicon Ceiling. You’ve made the old process cheaper, not better.

The New Procurement Playbook

Stop negotiating on cost-per-creative. Start negotiating on performance improvement.

Ask about data depth. How much media spend has trained the models? If the answer is “publicly available datasets,” walk away. You want platforms built on real brand performance.

Demand end-to-end integration. If the platform doesn’t connect directly to your media channels for generation, launch, measurement, and optimization, it’s creating a new silo, not solving the old one.

Measure creative lift, not creative volume. The wrong question is “how many ads can we generate?” The right question is “what’s the performance delta between AI-generated creative and our baseline?”

Require continuous learning loops. Static platforms that generate creative without feeding performance back into the system are yesterday’s technology. You need platforms that get smarter with every campaign, every impression, every conversion.

Insist on brand safety without bureaucracy. Leading platforms now automatically ensure AI content meets brand guidelines while maintaining velocity—protecting quality without killing speed.

Breaking Through the Ceiling

The brands that win with Gen AI won’t be the ones generating the most creative or saving the most money. They’ll be the ones that reunite creative and media into an intelligent system—where performance data continuously improves creative output, creative intelligence produces better ads, and better ads drive better business outcomes.

This means taking those production savings and reallocating them into working media. It means connecting your Gen AI platform directly to media channels for real-time optimization. It means measuring effectiveness, not just efficiency.

The technology exists. Platforms like Pencil have already processed billions in media spend, generating millions of assets that are predicted for performance before launch, automatically formatted for every platform, and continuously optimized based on real results.

The question isn’t whether to break through the Silicon Ceiling. It’s how quickly you can move before your competitors do.

Because here’s the final uncomfortable truth: your competitors aren’t stuck debating whether Gen AI saves money. They’ve already moved on to making more money.

About the author

Jack Smyth is Country Lead for Jellyfish + Brandtech Group in Australia

Jack Smyth is The Brandtech Group’s Country Lead for Australia. Our companies in Australia include OLIVER: the global leader for in-house agencies, Jellyfish; progressive digital services company powered by Gen AI, and Mobkoi; transforming mobile ads with innovative creative tech. Jack’s focus on Gen AI spans from mapping how AI models perceive brands, launching agents to predict social trends, and deploying our Gen AI creative optimisation platform – Pencil Pro. Before leading the Australia team he was CSO at Jellyfish, and previously held roles such as Creative Technology Officer at WPP’s Mindshare.