For years, one of the things keeping personalization in check was surprisingly simple: it was expensive.
A team could only build so many campaigns, storefronts, offers, and journeys. Before creating another version, someone had to decide whether it was worth creating at all.
AI is removing that constraint. Now we can generate more variations, for more customers, across more moments, at a speed that changes the economics of personalization entirely.
That’s a real advantage. More testing, more coverage, and more relevant experiences can create meaningful value. But abundance creates a different challenge. When you can make almost anything, choosing well becomes the hard part.
Most organizations are still responding with production metrics. More variants. Faster launches. More automated journeys. More output.
Those metrics tell us what the machine produced. They don’t tell us whether the customer experience got any better.
The advantage is shifting from how much you can produce to how well you can curate. That doesn’t mean putting a person in front of everything your AI has produced (which would simply recreate the bottleneck we just removed).
Instead, the goal isn’t simply keeping humans in the loop. It’s sharpening the judgment that loop contains.
Judgment isn’t some mysterious intuition that lives in the head of one talented marketer or merchant. It’s an organizational capability: the ability to make better choices about what deserves to exist, what feels right for the brand and customer, how individual experiences fit together, and where to draw the line.
Some of that judgment should remain with people. Some can be translated into principles and guardrails that AI operates within. The challenge is knowing which is which.
Think about photography.
When film was expensive, you thought about the photograph before pressing the shutter. The cost of production forced you to choose your shots carefully.
Now, the camera in your pocket lets you take hundreds of photos without thinking. Taking the picture is easy. The difficult part is finding the one good frame in a roll of hundreds, having the taste to recognize it, and the discipline to discard the rest.
AI is doing something similar to personalization. It is making production abundant. The hard part moves from creating the experience to deciding which experiences are worth having.
Production is becoming abundant. Good judgment isn’t.
The Judgment Layer
Organizations have spent years investing in the production layer. Now they need to intentionally design the layer that determines whether all that new output is actually worth putting in front of a customer.
The Judgment Layer comes down to five questions: What deserves to exist? Is it actually good? How will it land? Does it still feel like one brand? And where should we stop?
Consider what that looks like in a digital storefront.
Not long ago, a merchant might have hand-built a handful of homepage experiences, deciding what to feature and how products should come together. AI can increasingly assemble a different storefront for each visitor.
That doesn’t eliminate the merchant’s value. It changes where that value lives.
1. Prioritization: What Deserves to Exist?
AI can generate countless combinations of products, messages, and offers. The judgment is deciding which customer moments are worth personalizing in the first place.
When production is abundant, saying no becomes a capability.
2. Taste: Is It Actually Good?
A model can follow brand guidelines. Taste is knowing when something is technically on-brand but still doesn’t feel right.
The merchant’s role shifts from arranging every page to defining the standards the AI operates within.
3. Customer Empathy: How Will It Land?
Data can predict what a customer might do. Empathy considers how the experience will feel to the person receiving it, including when relevance starts to feel intrusive.
That’s a distinction AI signals alone can’t make for us.
4. Coherence: Does It Still Feel Like One Brand?
AI can optimize individual emails, product pages, and offers. Customers experience all of them as one brand.
The judgment is making sure thousands of personalized moments still add up to one coherent experience.
5. Trust: Where Should We Stop?
Optimization can lift almost any short-term metric. Trust is knowing when that lift comes at the customer’s expense.
Not everything that improves a short-term metric strengthens a customer relationship. Organizations need to decide where personalization crosses from helpful to manipulative or invasive, then build those boundaries into how AI operates.
Taken together, these five forms of judgment point to a different question for leaders. Not simply, “How far can we scale personalization?” but, “What judgment needs to guide it as we scale?”
Some judgment can be translated into principles and guardrails that help AI operate at scale. Other decisions require context, empathy, and taste that can’t simply be reduced to a rule. Getting that balance right requires designing for both, so greater scale doesn’t come at the expense of better choices.
That also changes what leaders should pay attention to. More content, more personalization, and faster execution tell us what AI made possible. They don’t tell us whether the resulting experience was more useful, more coherent, or more worthy of a customer’s trust.
As AI takes on more of the production, leaders need visibility into where judgment is being preserved, where it has been designed into the system, and where it may have quietly fallen away.
Every major technology shift changes what is scarce. For marketing, AI is making production abundant. The opportunity now is not simply to scale what we were already doing but to build the organizational judgment that scale requires.
The teams that stand out won’t necessarily be the ones that personalize the most. They’ll be the ones that know what deserves to be personalized, what good looks like, how it should feel, how it fits together, and where to stop.
AI can make the experience. Human judgment is what makes it worth having.
Want support putting this into practice? Let’s talk.
Aly Austin, Sr. Principal
