AI has changed the pace of everything.

You can feel it everywhere. Images. Headlines. Campaigns. Brand statements. Social posts. Web copy. Even the occasional freshly tailored neurosis. The machinery is humming, and some of it is genuinely useful. We’re not pretending otherwise. We’ve always had a soft spot for useful tools.

But after the novelty wears off, you start seeing the other side of it.

AI compresses the distance between blank page and presentable. And with fewer hours spent wrestling a first draft into something you’re not embarrassed to send, output gets easier. But the world doesn’t get clearer; it gets fuller. More polished language. More competent visuals. More things that look finished before anyone has really sat with them. The middle fills up fast.

This begins to reveal where the value lives.

In the old world, production had gravity. If something took more time, more rounds, more hands on deck, more expensive machinery, it was easy to believe that was where the magic was hiding. Turns out, not really. The tools matter. Of course they do. But the tools were never the whole show.

And when you can generate twenty versions before your coffee cools, you start to see that the real value was always in the call.

In the eye behind it. In the taste shaping it. In the instinct that says, no, not that one. In the restraint to leave the flashy thing on the bench because it doesn’t belong. In the small, painful, useful judgment that keeps a brand from flatlining.

That’s what becomes easier to see when everybody has access to the machinery.

A brand starts to feel dressed for the occasion, but not quite like itself. The copy is smooth, courteous, and perfectly capable of clearing a room. The visuals arrive pressed, coordinated, and expensive enough. Everything fits. Nothing belongs. No tension. No point of view. No pulse.

We’re seeing plenty of that already.

A lot of output. Very little authorship.

And the split is getting easier to spot. Heinz used AI to generate ketchup imagery, and the outputs kept drifting back toward Heinz. That became a sharp little argument for how distinctive the brand already was. The machine produced raw material; human judgment turned it into meaning. On the other side of the ledger, Coca-Cola’s AI-heavy holiday work landed with all the technical sheen in the world and still got read by plenty of people as cold, eerie, and off somehow. Fast. Scalable. Even well-scored in places. And yet, once it met actual human beings in the wild, the reaction was: unsettling, soulless, flat. That’s not a production problem. That’s a judgment problem. It’s what happens when nobody is really at the wheel.

And the broader pattern backs that up. Kantar found that AI-generated ads can land anywhere on the spectrum, but the ones where AI use is more seamless perform better, with more than 40% landing in the top tier for branded cut-through. The obvious stuff tends to fare worse. Same with branding. When the model is left to invent the look and feel on its own, without being trained into the brand’s tone, assets, and logic, the result is weaker branding on average. Which is another way of saying: the problem is not that AI makes things fake. The problem is that it makes it very easy to make things generic, and generic is what people feel first.

Which is worth paying attention to.

The smarter brands are starting to respond accordingly. TD’s new “More Human” platform is built on the idea that as digital systems and AI make life more efficient, they can also make experiences feel colder and less personal. So the job is to use the technology without letting the humanity drain out of the thing. Dove has drawn an even harder line, publicly pledging to keep its work free of digital and AI distortion and to never use AI imagery in place of real women. Different categories, different thresholds. Same underlying read: once trust, identity, memory, appetite, or belonging are involved, synthetic polish gets expensive fast.

This is where teams can get themselves in trouble.

Speed starts overruling instinct. Metrics start talking people into work their own taste might have rejected on a better day. Infinite options create the illusion of abundance when what you’ve really got is more ways to be mediocre at scale. And if the process itself hasn’t changed to account for that, technically correct, aesthetically smooth, strategically vague, emotionally weightless slop slips through. So if you’re not careful, you don’t just get faster. You get easier to forget.

That’s the pressure point at this moment.

As production gets easier, discernment is not just garnish.

Taste matters more now because there’s more to sift through. Intuition matters more because the line between strong and generic can vanish in a blink. Judgment matters more because the machine will keep offering options long after somebody in the room should have said, enough, kill it.

And judgment — real judgment — is not some dry executive function. It’s not a checklist. It’s feel. Proportion. Timing. It’s knowing when a line needs tightening, when a visual system is trying too hard, when a homepage is overexplaining, when a campaign has crossed the line from distinct into strained. It’s knowing when something has life in it.

That kind of read doesn’t come from nowhere.

It comes from experience. From paying attention. From getting your ass kicked a few times. From making a lot of things over a long stretch of time and learning, sometimes painfully, what holds up and what just looks good for five minutes. There’s no substitute for experience there. None.

After nearly 40 years, that’s still the part we trust most.

Not because we’re sentimental about craft. And not because we think every new technology is suspect for some romantic old-world reason. We’ve always moved with the tools — often early. But the work still has to clear the same bar for a damn good reason. It has to feel considered. It has to carry conviction. It has to sound like somebody meant it.

Otherwise, what are we doing?