Brennan McGrath is a Partner at Combo, where he helps brands grow by blending cultural insight with creative ambition. Guided by a deep belief in empathy, he approaches every project with a relentless curiosity about what drives people — what moves them, inspires them, and ultimately shapes their behavior. It’s a principle that has defined his work across a wide range of clients, including Diageo, Target, and Tishman Speyer. A former architect and lifelong builder, Brennan brings a hands-on, systems-level perspective to branding. He’s spent over a decade running creative agencies, evolving his independent practice from Verde to what is now Combo. Outside the studio, Brennan is a long-distance runner, outdoor enthusiast, and someone who genuinely enjoys making things — from physical space to provocative ideas. His craft is rooted in a belief that the best ideas come from listening deeply and building with care.
***
There’s this growing idea that as AI makes everything, taste becomes the differentiator.
That’s probably true. But the way we’re defining taste right now is off.
Most people talk about taste like it’s refinement. Curation. Knowing what’s good. But that version of taste is exactly what AI is getting very good at replicating. It can absorb patterns, learn what’s been validated, and generate something that looks right.
That’s not taste. That’s criteria.
Take something like Claude Design. You can describe what you want, from a landing page to a pitch deck to a campaign visual, and it will generate something polished, on-brand, and structurally sound in seconds. No design background required.
That’s incredible! It’s also the point.
Because what Claude Design is actually doing is operationalizing what we’ve been calling “taste” for years: alignment to patterns, consistency with a system, adherence to what’s already been validated.
It’s very good at making something that looks “right”. But it doesn’t have the ability to believe in something that isn’t right yet.
AI doesn’t have taste. Not in any real sense. It doesn’t have a tongue that can like and dislike flavors. It doesn’t have a heart that can break. It doesn’t have the relationship with consciousness and survival necessary to understand why one thing is better than another. It can be told what’s good, but being told isn’t the same as knowing.
So when people say “AI doesn’t have taste,” they’re not really defending some precious creative instinct. They’re reacting to something else: the realization that we’ve spent a long time using curation as a proxy for relevance. And now that proxy is being automated.
The truth is, we’ve already been living in a world where taste is flattened.
Every H&M plays the same music. Every city has the same Scandinavian-looking coffee shop. Algorithmic media has pushed us into a place where everything is oddly unique and uniform at the same time. We’re surrounded by signals of taste that all point in the same direction.
AI doesn’t break that system. It accelerates it.
As AI makes it easier to produce competent, on-brand work at scale, “good” becomes the baseline. The middle flattens. There are more options, more outputs, more things that look right.
So the advantage shifts. Not to the people or brands who can make the best-looking thing, but to the ones who can make the most unexpected decision.
Because real taste isn’t about knowing what’s good. It’s the ability to see value in something underappreciated. To look at something overlooked, underdeveloped, or even a little wrong — and understand what it could become.
It’s the person who picks up a necklace at a flea market because they love it, not because they recognize it. Years later, they find out it’s an original Alexander Calder. That instinct, the ability to assign value before it’s validated, is taste.
AI can tell you what something resembles. It can map it to patterns. But it can’t love the curious genius of the thing itself.
More importantly, it can’t misunderstand.
Humans have an innate ability to misunderstand things and embrace contradiction. Those two qualities — misunderstanding and contradiction — are the source of so much that’s actually new. Every good joke. A lot of new music. Every counterintuitive business.
Without them, you don’t get a Liquid Death. You don’t get a Cowboy Carter. You don’t get ideas that don’t immediately make sense, but end up reshaping categories.
That’s the gap.
As AI gets better at replicating “good taste,” the brands that win won’t be the most tasteful in the traditional sense. They’ll be the ones willing to choose something that doesn’t immediately read as correct. The ones that can hold a little contradiction without smoothing it out. The ones that can believe in something before it makes sense.
We’re not entering a taste economy. We’re entering a belief economy. And most brands are still optimizing for approval.
