Why $1M Ads Still Start With People, Not Prompts
Why It Matters
Right now, many people are excited about automation, and it's tempting to think that more AI results mean more success. However, Harry's data shows otherwise: simply increasing volume without creative ideas only leads to more average ads created more quickly. For media buyers and creative strategists, this is a helpful reminder. Focus on safeguarding the human touch in research and idea generation, and let AI support the execution. That's what makes the difference between an ad that disappears in a few days and one that can grow and perform for months.
Who This Is For
Created specifically for Founders members who are working on creative strategies or managing accounts with real spend on Meta. Especially if you're feeling the pressure to "automate everything" and worry about whether you're cutting corners. If your team is relying on AI for generating ideas rather than putting them into action, this is a helpful reality check.
You’ve probably already heard the hype that is “AI is revolutionizing ad creative!” “Just automate everything and watch the sales roll in!”
However, the ads that truly drive $1 million or more in spend and scale brands are still crafted the traditional way: meticulously, through hands-on work, thorough research, and a unique spark of human creativity.
In a recent episode of AI DTC WTF, I sat down with Will Sartorius and Harry Delmege, co-founder of Roly Poly Digital, to get into the real role of AI in ad creative, where it genuinely helps, and where it quietly kills your hit rate. Harry’s team is best known for “yapper” ads, the long-form, story-driven copy that has scaled past $1M in spend. They rely heavily on AI for execution but almost never for ideation. This conversation is a must-read for anyone who wants to understand the actual mechanics behind winning digital ads in 2026.
The AI Hype Machine: Where It Helps, Where It Hurts
Harry isn’t anti-AI – he uses it, but he understands its limits. He says that AI can help execute at scale by formatting, rewriting, mining reviews, and speeding up grunt work. But when it comes to ideation, the magic still happens with humans.
“The bottleneck isn’t execution anymore,” Harry says. “It’s ideation. AI just can’t come up with the ad angles that win.” You can automate the process, but you can’t automate the spark. The best ad ideas are the ones that unlock new markets, and these tell compelling stories, drive scale, and come from manual research and human creativity.
Building AI Skills for Execution And Not Ideation
Harry’s team has built out a toolkit of AI skills for different ad formats: yapper ads, VSLs, native long-form, ad analyzers, and more. They’re stacking these skills on top of what’s already worked: past winners, team notes, and the basics of great copywriting. It speeds things up and keeps the quality tight, but the foundation remains in human creativity.
“We only take the winners and bake their essence into the skill,” Harry explains. “But the initial spark, the angle, the story, the hook, is always manual.”
Manual Research Still Beats AI Research
One of the episode's biggest takeaways is the importance of manual research. Harry runs a monthly “manual research bounty”, where the team posts their best findings, and if someone uncovers a major unlock for an account, they get a bonus. Harry says he does this because AI just isn’t good enough at finding the customer stories that make ads work.
Harry’s even tried letting AI loose on Reddit to dig up stories, but the results were underwhelming. “AI just skims everything,” he says. “It gives you generic summaries instead of the deep, specific stories you need.” If you want that angle nobody else has, you’ve still got to roll up your sleeves and dive into the deep research yourself.
AI UGC Fatigues Faster Than Real UGC
AI-generated UGC can help create rapid iterations and test new formats, but Harry has noticed that AI UGC dies faster than real UGC, even when it works well at first.
“We’ve seen AI UGC crush for a few days, then fall off a cliff,” he says. “Real UGC, with real people and real stories, has legs. It can stay at the top of the ad account for months.” AI UGC is great for short-term wins, but the lasting impact comes from authentic content.
Founders members regularly exchange operator details such as manual research bounties, human QA, and ideation frameworks. Want to be part of the community?
Where AI Creative Genuinely Wins
AI creative so far is unbeatable with formats like claymation, animated Pixar-style ads, wild B-roll you could never shoot in real life, and impossible-to-shoot-within-a-reasonable-budget visuals.
Harry shares an example: for a pet brand, his team used AI to create a claymation-style ad with a talking dog. “Three years ago, this would have been impossible for most advertisers,” he says. “Now, you can generate it in a couple of hours or workdays.”
AI blows open the doors on what’s possible, letting brands try creative formats and wild ideas that would have been way too expensive or impossible before.
What a $200K/Month Ad Account Looks Like in 2026
When you’re spending $200K a month or more, the best ad accounts are a mix of creative diversity, such as founder content, yapper ads, performance UGC, AI-generated statics, and more. The goal isn’t to have formats fighting for budget, but to have every format spend and scale together.
Meta wants diversity, and AI helps deliver it, but diversity alone isn’t enough – quality still matters. The winning accounts are the ones that combine manual ideation and research with AI-powered execution and testing.
Quality is the Name of the Game
There’s an industry myth that volume wins: launch hundreds of new concepts a week, and you’ll find the next big thing. But Harry says "it’s about the quality of what you’re creating.”
AI lets you crank out a ton of volume, but that doesn’t mean better results. The real needle-mover is making a handful of ads so good that they stay live and scaling for months and keep compounding results.
The Pendulum Swings Back to Fundamentals
In the early days of AI, the pendulum swung toward automation: one-shot scripts, infinite ad launches, skipping the hard work. Now, the industry is swinging back to fundamentals: manual QA, deep research, and copywriting craft.
Harry emphasizes the importance of human QA: “Ads that pop off are the ones where you add tiny details, make small changes, and add nuance. The best ads go from a six out of ten to a nine out of ten because of human touch.”
Brand owners are firing agencies over “AI slop” ads that feel generic, lack story, and don’t connect. The demand for quality is rising, and agencies are being forced to bring back manual skills.
The One Quality to Hire Creative Strategists For
When hiring creative strategists, Harry looks for one thing: willingness to do the hard work. Experience matters less than hunger and curiosity. “I want people who are still willing to learn, break down winning ads, QA scripts, and refine their craft,” he says.
Many candidates lean too hard on AI, sending in copy that’s good enough but not great. The best strategists go the extra mile; they’re hungry, curious, and always chasing excellence.
Keeping Teams Sharp: Manual Research Bounties
To keep his team’s skills sharp, Harry runs monthly manual research challenges. Financial incentives make it worthwhile, but the real goal is to maintain the habit of deep research and manual ideation.
“It’s harder now,” he admits. “AI makes it easy to skip the manual work. But if you want to keep your tools sharp, you have to do it.”
Incentivizing the right behavior is key. Rewarding manual research and high-quality manual ads drives better results and keeps the team from getting too AI-dependent.
Brand Governance and AI: Finding the Middle Ground
Some brands are loose and let agencies run whatever they want. Others have strict governance and can’t even get AI animation approved. Harry sees this as the “death of advertising”, brands missing trends because approval processes are too slow.
The sweet spot is somewhere in the middle: enough freedom to experiment and push boundaries, but enough governance to keep brand standards tight. As AI creative gets more common, brands will need to get clear on their policies and approval processes.
New episode, new operator, same real talk on what AI can and can't do for your ads.
The Future: Bringing Imagination to Life with AI
Looking ahead, Harry’s betting the winners will be the ones who use AI to bring their wildest ideas to life. Efficiency will keep getting better, but creativity is still the secret sauce.
“The people who crush it will be the ones who keep their imagination and creativity alive,” he says. “AI lets you create wild things, but you need strong scripts and stories. Advertising hasn’t changed. We’re still telling stories, just with new mediums.”
Manual research, deep ideation, and human QA aren’t going anywhere. AI will help with execution, reporting, and all the grunt work, but that creative spark is still 100% human.
Avoiding the AI Trap
One of the biggest traps is thinking that AI-generated content is enough. Creative strategists may know this, but many leaders don’t. The call to action is to not just push the volume button. Don’t settle for personas or generic scripts. Push for quality, story, and human creativity.
The best ads are built on manual research, deep ideation, and relentless QA. AI is a powerful tool, but it’s not a replacement for hard work.
Conclusion: The Real Edge in Digital Advertising
As AI keeps evolving, the temptation to automate everything is only going to get stronger. But the real edge (the ads that spend $1M and actually scale brands) is still built by hand. Manual research, human creativity, and quality QA are what move the needle.
AI is a teammate, not a replacement. Use it for execution, review mining, and rapid iteration. But keep your creative process rooted in manual ideation and storytelling.
If you want to build ads that last, invest in your team, reward manual research, and hire strategists who are hungry to learn. The future belongs to those who blend technology with creativity and never lose sight of the fundamentals.
Check out the full AI DTC WTF episode:
Note: This blog was written by a human, but AI helped draft the outline and synthesize ideas.
