Every marketing leader I talk to is obsessed with AI-powered personalization. The promise sounds amazing-serve the perfect message to the perfect person at the perfect moment. Who wouldn’t want that?
But here’s what nobody is saying out loud: we’ve crossed a line. That hyper-targeted ad that knows exactly what you were looking at 20 minutes ago? It doesn’t feel helpful anymore. It feels like surveillance.
At Sagum, we’ve been running campaigns across every major platform for over a decade. We’ve spent millions on Facebook, TikTok, Google, and YouTube. And the data has taught us something that most AI engineers don’t want to hear: the smartest personalization strategy isn’t showing people what they want. It’s showing them what they need in a way that actually feels human.
Here’s the real problem and what to do about it.
The Three Traps of Over-Personalization
Before we talk about solutions, we need to understand what’s actually breaking. In my experience, AI personalization engines create three predictable failures.
The Mirror Trap
When your AI only shows people content that mirrors their existing behavior, you’re killing discovery. Think about it this way: if the person who buys a Rolex for their 50th birthday only saw watch ads when they searched for watches, they’d never buy anything special. They needed to discover it long before they knew they wanted it.
AI engines don’t know how to create desire. They only know how to harvest existing intent. That’s a fundamental limitation.
The Data Creep
This one is obvious but ignored. When someone sees an ad that says “Hey, I noticed you looked at these blue shoes 47 minutes ago,” the reaction isn’t “wow, great service.” The reaction is “how do they know that?“
Trust takes years to build and seconds to destroy. One aggressive personalization play can make a customer block your ads forever.
The Zero-Sum Game
Most AI engines optimize for short-term return on ad spend. They strip away storytelling in favor of hyper-targeted product pitches. You win the click, but you lose the customer. They buy once, feel weird about it, and never come back. Why? Because the brand never meant anything to them. You optimized the transaction and sacrificed the relationship.
The Fix: The 70/30 Rule
After years of trial and error, we’ve developed a framework that works. We call it the 70/30 Rule. It’s simple but powerful.
70% Context Over Identity
Stop asking “who is this person?” Start asking “what is this moment?”
- What platform are they on? Instagram Stories feels different than YouTube pre-roll.
- What’s their state of mind? Are they browsing for fun or searching with intent?
- What format works here? A 15-second vertical video isn’t the same as a search ad.
This is contextual intelligence. It doesn’t require knowing someone’s browsing history. It only requires understanding the moment they’re in right now.
30% Intent Over Identity
Here’s the critical distinction that changes everything:
- Don’t say: “I saw you looked at this product.” (Creepy, intrusive, bad.)
- Do say: “People who care about quality also love this.” (Helpful, expert, good.)
The difference seems small, but it’s the difference between sounding like a stalker and sounding like a trusted advisor.
How to Actually Execute This
Theory is easy. Execution is hard. Here’s the tactical playbook.
- Kill the perfect profile. Stop trying to build 100,000 micro-audiences. It’s a waste of energy. Instead, define 3-4 broad mindsets-The Skeptic, The Hobbyist, The Professional-and let AI optimize the creative format within those groups.
- Use generative gaps. Leave space for the human to connect. Instead of hyper-personalized headlines, use bold universal truths: “Most companies charge for this. We don’t.” Let the AI personalize the button color, not the core message.
- Shift from behavioral to aspirational. Stop asking what someone did 5 minutes ago. Start asking who they want to become. Bad AI says: “You left items in your cart.” Good AI says: “You deserve gear that matches your ambition.”
The Anti-AI Advantage
Here’s the uncomfortable truth that will separate winning agencies from everyone else: the best marketing teams won’t be the ones with the fanciest machine learning models. They’ll be the ones who know when to ignore the machine’s recommendations and trust human judgment instead.
That’s why we limit our client roster at Sagum. We can’t afford to be distracted by every shiny new AI tool. We need to stay focused on what actually works: gaining traction, hitting real targets, and building brands that last.
AI personalization is a powerful tool. But a tool without a skilled operator is just noise.
One Question to Ask Yourself
Before you deploy your next personalized campaign, ask this question:
“Would I be okay if a competitor showed this exact same ad to my customer?”
If the answer is yes, you’ve built a commodity. Your personalization strategy is replaceable.
If the answer is no, you’ve built a brand. And that’s something worth protecting.
