Open ten pitch decks from ten different startups this year and you'll see the same brand: a rounded sans-serif, a gradient somewhere between violet and teal, a blob shape doing duty as a "brand asset," and a wordmark that could belong to any of them. This isn't a coincidence. It's what happens when an entire industry starts designing with the same tools, trained on the same reference pool, optimizing for the same statistically "good" output.
We've been saying for a couple of years that AI would democratize design. It has. What nobody wanted to say out loud is that democratizing the process was always going to homogenize the output. When everyone has access to the same generative tools tuned toward the same aesthetic consensus, "good design" stops being a differentiator and starts being table stakes. Which means it stops being brand equity at all.
The tools aren't the problem. The brief is.
I don't think this is an argument against AI in the studio. We use it constantly: for moodboards, for exploring typographic pairings faster, for mockups we'd never have bothered making by hand. The efficiency gains are real and I'm not interested in pretending otherwise.
The problem is upstream of the tool. Most brand work that comes out generic didn't get generic because someone typed a prompt into a model. It got generic because the strategic brief was thin to begin with ("modern, clean, trustworthy, a little playful"), and a thin brief produces a thin brand whether a human or a machine executes it. AI just executes thin briefs faster and at greater scale, which is why we're seeing sameness compound across an entire funding cycle of startups rather than one company at a time.
The studios getting this right aren't the ones avoiding AI. They're the ones who've gotten stricter about what goes into the brief before any tool touches it: a specific point of view, a competitor a client is willing to look nothing like, an opinion about what the brand should never be. Specificity is the only thing that resists flattening. Vagueness is exactly what gets smoothed toward the mean.
What "distinctive" has to mean now
There's a useful test we've started applying internally: could this identity survive being run through the same generative process that made everyone else's brand? If a logo, palette, or voice would come out the other side indistinguishable from a competitor's, it wasn't distinctive to begin with. The AI just made that visible faster than a slow market cycle used to.
That means leaning harder into the things that are genuinely hard to average out: a strong point of view in the copy voice, an idiosyncratic color choice that a committee would have talked someone out of, a system built around a real asset (a founder's handwriting, a physical material, an actual photograph) rather than a generated abstraction. Distinctiveness increasingly means friction. Something a model would have smoothed over if a human hadn't insisted on keeping it in.
For clients, this changes what they should be paying a studio for. It's not execution speed anymore. That's commoditized. It's judgment: knowing which weird, specific, slightly uncomfortable choice is the one worth defending against the pull toward consensus. That's the actual service. Everything else is increasingly a feature of the software.
The brands that will feel dated in two years aren't the ones that used AI. They're the ones that let AI's aesthetic center of gravity make their decisions for them.
