Tanmay Ratnaparkhe, Co-Founder, Predis.ai, which uses AI to help brands scale ads, ad videos and social content without losing their voice.
The marketing teams I speak to all share the same point: They can now produce more advertisements in less time than they could a year ago. Work that once required a week of filming and revisions can now be done in an afternoon with a few prompts. The output figures look impressive. Yet when I ask whether the work is actually performing better, I usually get a pause.
What makes the pause so interesting is that for the past two years production speed has been our aim, and I now believe that we should have chosen a different target. Once anyone can produce 100 variations before lunch, speed stops being an advantage and becomes the basic standard. When everyone has the same basic standard, however, it no longer helps you to stand out.
Speed Doesn’t Win Anymore
For nearly the entire history of advertising, the difficult part was production. Making advertisements was a slow, costly process, so the ability to produce more was important. That is no longer the case. The tools that enable my team to work quickly are generally the same ones all competitors have. Once an entire category can produce at the same speed, speed stops being the deciding factor.
The unpleasant thing is that the capability everyone has rushed to develop has little defensive value; if a competitor can match your output by Friday, then output was never really your advantage. Experience shows that teams that still assess themselves primarily on volume are most surprised when their performance levels off.
Why Faster Production Drifts Toward Sameness
A more fundamental reason the race tends everybody toward the middle is that this tendency is inherent in how the tools function. Generative models are trained to generate what is most statistically likely—the most probable following line, the most probable next arrangement. At their default settings, they move toward the average of all the examples they have already seen. This is not merely a guess. A meta-analysis of creative tasks in 2026 found a small but statistically significant effect of homogenization when people used generative AI, compared with when they didn’t. When given similar briefs and using similar tools, people produced similar results.
I can see this taking place in creative reviews. The initial ideas all seem original, but by the 20th they start to blend together—merely competent, fashionable and forgettable. The speed that once felt liberating gradually creates a house style no one actually chose. Faster iteration doesn’t help you avoid this problem; it only gets you there faster.
The Part That’s Actually Scarce Is Judgment
If production isn’t the bottleneck, the scarce resource has shifted to judging which ideas are worth showing people. It’s now easy to come up with 50 options, but the difficulty lies in deciding which of them should actually receive a real budget, and having the ability to justify that decision. No tool will do that for you.
On my teams, the most valuable person isn’t the one who produces the most assets; it’s the one who, when faced with a multitude of options, can give a genuine reason for saying none of them feel like our brand. That sort of judgment—a clear understanding of what the brand is and what it isn’t—is something automation can’t provide. In my experience, it’s also what distinguishes work that gets carried forward from work that is simply scrolled past.
Put The Hours Back Into The Idea
The solution is not to reduce the speed. That speed is genuine and has a useful purpose. The issue then becomes what you choose to spend it on: You can use it to produce more of the same, or you can use it to make genuine attempts at being distinctive and discard the ones that fail.
As a result, the way we work changes. Rather than asking the tool for a hundred usable ads, I use the time that is saved to improve the brief, clarify the point of view and set a higher standard for what is good enough to run. You should treat generation as the cheap part and judgment as the expensive one, since that better reflects current costs. The teams I observe advancing aren’t the ones who produce the fastest. They are the ones most ready to discard work.
The competition to move faster will keep accelerating, and most of the industry will keep joining in. However, the result of this competition is the middle ground—where nothing really stands out. The more valuable race, and the one fewer teams take part in, is simply being worth noticing at all. In that race, the winner is not the person who produces the most; instead, the winner is the one who knows what is worth producing.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?