Between the brief and the brand, there's a gap
Every agency knows the story: the client wants "something creative that converts." The brief closes. The team works. And in a world where AI can generate a complete ad spot in a few hours, the temptation is real. Why wait two weeks for a concept when you can have a storyboard at 11 AM and a finished video by lunch?
It works on paper. Click-through rate, completion rate, cost per view, even engagement. Green numbers fill the dashboard. The client is happy. And yet, something is missing. Nothing quantifiable. Something you feel when you see an ad that looks right, says the right things, but has no soul.
2026 is the year major brands started acknowledging this publicly. Coca-Cola, after two seasons of AI-generated holiday ads, returned to traditional creative, shot with real actors and a professional director. Not because the AI failed technically. But because the audience reacted differently than the numbers suggested.
And they're not alone. The trend is broader. More and more Fortune 500 companies are revising their AI creative strategies. They're not abandoning AI. They're moving it from center stage to backstage. It's a subtle but fundamental shift in how the industry views the relationship between technology and creativity.
The metrics say yes. The gut says no.
There's a paradox we're seeing more and more in the industry. AI-generated ads pass performance tests. A 2026 Kantar report shows that AI-created advertising can match or even exceed traditional benchmarks on awareness and recall. But the same report reveals something subtler: the "emotional resonance" score is 40% lower.
People watch. People understand the message. But people feel nothing.
In a market saturated with content (500 million stories daily on Instagram alone), a brand's real advantage isn't visibility. It's the ability to be remembered. To trigger an emotional response. We've written before about how the ads that stick aren't the ones that sell directly, but those that build associations. AI isn't good at associations (yet). It's good at patterns.
Patterns are useful in targeting, segmentation, media optimization. But in an ad, the pattern is exactly what makes you invisible. Audiences don't react to efficiency. They react to surprise, to authenticity, to that moment of "that's exactly what I was feeling." And that moment comes from human experience, not training data.
The "almost right" effect
Consumers in 2026 have an AI radar they didn't have two years ago. It's not about visual artifacts (extra fingers, odd textures). It's something subtler: a uniformity in tone, a predictability in structure, an absence of the imperfection that makes human content feel human.
A 2026 Ipsos survey shows that 62% of consumers say they would trust a brand less if they found out its ads were AI-generated. Not because the AI produces something bad. But because it signals something else: "This brand didn't invest enough in its relationship with me to have a human think about me."
It's what we call, internally at difrnt., the "almost right" effect. Everything looks correct. Nothing is wrong. But nothing stops you from scrolling. Nothing makes you message a friend with "check this out." In a world where people receive hundreds of advertising messages daily, "almost right" is the equivalent of invisible.
The problem becomes sharper for brands with history. A tech startup can use AI creative without major consequences because it doesn't have an emotional heritage to protect. But for a Coca-Cola, a Nike, or even a local brand with a loyal community, every ad is an implicit promise. "We put care into this. We made it for you." When the audience feels that care is missing, trust erodes, even if the product stays the same.
What we actually do
At difrnt., we're not anti-AI. Far from it. We use AI every day. But AI's place is in the process behind the scenes, not in the output upfront.
We use AI for research. For audience analysis. For headline testing. For reporting automation. For brainstorming creative directions, as a starting point, not a destination. AI helps us understand faster. We decide what and how to communicate.
In practice, our workflow looks like this: AI processes data and trends. The creative team interprets. The strategist validates against business objectives. Then the copywriter writes. A human who senses the right tone, who knows when text sounds corporate and when it sounds like a conversation. We've written about how the cost of creative production dropped dramatically with AI. That's true. But reducing cost doesn't mean eliminating the human. It means the human can focus on what matters: the idea, not the mechanical execution.
A simple example: for a recent performance campaign, we generated 40 ad copy variants with AI. From those 40, the team selected 6 that had something extra, a nuance, a tension, an unexpected angle. We rewrote them manually. The 6 final variants outperformed the original 40 by 35%. Not because the AI was wrong. But because the human filter was missing.
Where the industry is heading
Our prediction: by 2027, no serious brand will promote the fact that its ads are "AI-generated." On the contrary, "human-made" will become a differentiator. Just as "organic" became relevant in food and "handcrafted" in artisan goods, "human-made" will become a quality signal in advertising.
This doesn't mean AI disappears from advertising. It means AI finds its real place: in the processes behind the campaign, not in the output the consumer sees. A brand's creative culture remains a differentiator no technology can replicate.
The question we ask our clients isn't "do you want to use AI?" It's "where do you want the audience to feel your hand, and where can you automate without losing anything?" The answer is almost always the same: AI in data, AI in optimization, AI in workflow. Humans in the message, humans in the concept, humans in the tone of voice.
The brands that will win aren't those that produce more content. They're those that produce something worth remembering. And that, for now, still requires a human in the equation.



