Here’s something nobody wants to admit: the AI “best practices” everyone’s following are turning marketing into vanilla pudding-safe, predictable, and completely forgettable.
I’ve been in this business long enough to recognize patterns. Right now, I’m watching agencies and brands race toward the same cliff, convinced they’re innovating when they’re actually commodifying themselves. Everyone’s using the same tools, following the same playbooks, and wondering why their marketing feels like everyone else’s.
The irony is almost funny. AI was supposed to give us superpowers. Instead, it’s given us sameness at scale.
The Question Nobody’s Asking
Here’s what keeps me up at night: if every brand in your category is using the same AI tools with similar prompts to solve similar problems, how does the output create any competitive advantage?
It doesn’t.
Think about your Instagram feed right now. How many captions have that telltale AI rhythm? That oddly formal-yet-casual tone that sounds like it was written by a very polite robot? How many emails in your inbox could’ve been written by literally any brand in that category?
We’ve created a digital landscape where everything is optimized and nothing is memorable.
What Actually Works: The Three Zones
After spending millions testing campaigns across every major platform, I’ve developed a framework that actually makes sense of when AI helps versus when it hurts. Think of your marketing in three zones:
Zone 1: Efficiency Theater
This is where most brands dump their AI budget-automating stuff that didn’t need automating. Social captions. Email subject lines. Blog outlines. All the tasks that were already pretty efficient.
The real problem? These touchpoints shape how customers perceive you. When you automate them, you’re optimizing for small efficiency gains while sacrificing your distinctive voice. It’s a terrible trade.
My rule: if it touches customer perception, approach AI with extreme caution. Maybe don’t use it at all.
Zone 2: Intelligence Multiplication
This is where things get interesting. Instead of replacing human thinking, AI multiplies it.
For example, don’t ask AI to write your customer personas. That’s lazy and you’ll get generic garbage. Instead, feed it thousands of customer reviews, support tickets, and social comments, then have it surface patterns. Patterns a human could never spot manually. Then-and this is critical-YOU interpret those patterns with your strategic brain.
Same with creative testing. Let AI generate 100 variations of your message so you can see patterns in what might resonate. But the human still crafts the final winner, informed by unprecedented input.
This is AI as a thinking partner, not a replacement.
Zone 3: Competitive Moats
Almost nobody plays in this zone, which is exactly why you should.
This is where you use AI to build capabilities competitors can’t easily replicate. Custom models trained on your proprietary data. Systems that get smarter the more you use them, creating compound advantages over time.
If you’re running campaigns across multiple platforms, AI can analyze performance patterns specific to your brand, your audience, your creative. It builds a self-improving system that becomes more valuable with scale. Your competitors using off-the-shelf tools can’t touch this.
Five Principles That Changed How We Work
1. Leave Human Fingerprints Everywhere
Every customer-facing piece of content should have something unmistakably human in it. An unexpected analogy. A specific story. A controversial take that only someone with real expertise would share.
I’m not saying don’t use AI. I’m saying if you do, transform the output into something that couldn’t have been generated. When we create content, we make sure it reflects our actual methodology and client experiences-things AI can’t fake because they’re ours alone.
2. Build on Proprietary Data
Generic AI tools trained on public data give you generic outputs. Revolutionary insight, right?
The real opportunity is training models on your specific data. Your customer purchase patterns. Your seasonal trends. Your product affinity maps. This creates insights and recommendations your competitors literally cannot replicate because they don’t have your data.
3. Make It Weird Fast
Here’s my favorite metric that doesn’t exist in any dashboard: speed-to-weird. How quickly can you take AI output and make it specific, unexpected, or distinctively yours?
If your AI-generated ad copy could work for any competitor, you’ve failed. The goal isn’t optimal-it’s distinctive. When we test ad variations, we push past AI suggestions and combine its structural insights with our brand’s specific worldview. “Efficient” becomes “ruthlessly efficient.” “Customer-focused” becomes “obsessively customer-focused.”
Small shifts. Massive difference in how it lands.
4. Be Strategic About Transparency
Here’s the thing about disclosing AI use-it’s not just ethics, it’s strategy. How people feel about AI in different contexts varies wildly.
Internal tools? Nobody cares, don’t mention it. Customer service bots? Be upfront, but frame it right: “Our AI handles routine questions instantly so our humans can focus on complex problems.” That positioning matters.
Strategic content and decisions? Always human-led. Always.
5. Break the Benchmarks
Most AI implementations aim to hit benchmarks. I want you to break them.
If your AI analysis says email open rates should be 22%, don’t celebrate 23%. Ask what assumptions the AI made that you could deliberately violate. Maybe your audience would love radically longer subject lines. Or emails sent at 2am. Or content that breaks every category norm.
When our campaigns perform exactly as AI predicts, we know we’re playing the same game as everyone else. We win when we find the approaches AI didn’t see coming.
How to Actually Implement This
Strategy without execution is just therapy. Here’s the framework:
Step 1: Run an Authenticity Audit
Before you implement any new AI tool, ask one question: “Will this make us sound more like ourselves or more like everyone else?”
Map every customer touchpoint. Mark the ones that need distinctive voice, unexpected perspective, or deep expertise. Those are either AI-free zones or places where you heavily modify any AI output.
Step 2: Build Your Intelligence Layer
Generic AI tools are table stakes now. They’re not your strategy.
Your strategy is building proprietary intelligence:
- First-party data infrastructure that captures what matters
- Custom datasets that train models on your reality
- Performance patterns unique to your brand
- Audience insights competitors don’t have
This is how AI becomes a moat instead of a commodity.
Step 3: Define Collaboration Protocols
Get specific about where AI and humans intersect:
- AI generates options → Humans select and transform them
- Humans set strategy → AI executes and optimizes
- AI surfaces patterns → Humans interpret with context
For us, this means AI handles bid optimization across platforms while humans own creative decisions and strategy. The AI makes thousands of micro-decisions. Humans make the few decisions that actually matter.
Step 4: Measure Differentiation, Not Just Performance
AI makes it easy to measure clicks, engagement, conversions. That’s the problem-you optimize for what’s easy to measure instead of what matters.
Start tracking:
- Brand recall compared to category averages
- Message uniqueness (does your copy sound like you or like everyone?)
- Customer feedback on personalization (helpful or creepy?)
- Creative fatigue rates
- Voice consistency over time
Step 5: Systematize Weirdness
This sounds contradictory, but it works. Create processes that force you to deviate from AI recommendations:
- Every AI-generated campaign needs one “human override” element
- Weekly meetings to discuss “what would AI never suggest?”
- Monthly creative sprints that completely ignore AI insights
This prevents the slow, invisible drift toward sounding like everyone else.
What This Looks Like Platform by Platform
Facebook and Instagram
The trap is using AI to generate endless variations that all perform “okay.” You end up with 50 ads that get a 1.2% CTR instead of finding the one ad that gets 4%.
Better approach: Use AI to identify your top 5% of creative patterns based on your account data (not industry benchmarks). Then have humans create bold variations that push those patterns to extremes.
After years scaling Facebook campaigns, the pattern is clear-AI identifies that audiences respond to specificity, then humans create ads so specific they feel almost personalized even though they’re not. That’s where the wins live.
TikTok
Do not-I repeat, do not-use AI to analyze trends and copy them. TikTok’s algorithm punishes imitation ruthlessly.
Instead, use AI to understand trend structures. Why do certain formats work? What’s the underlying pattern? Then create original content that uses those structures with your perspective baked in.
We’ve spent over $2 million on TikTok ads, and the data is brutal: AI-generated content gets flagged as inauthentic by both the algorithm and real humans. But AI that analyzes why human content works? That intel is gold.
Google Ads
Google’s AI bidding is genuinely good. Use it. But don’t let automation run your entire strategy.
Let AI handle bids and budget allocation. Humans should own messaging strategy, audience insights, and the creative testing roadmap. From Search to Shopping to Discovery, AI optimizes but doesn’t innovate. That’s your job.
YouTube
AI-optimized thumbnails and titles all start looking the same. Shocking face. Yellow text. Red arrow. You’ve seen it a thousand times because AI says it works.
It does work-until everyone does it. Then it’s invisible.
Use AI to understand what drives clicks in your category, then break those patterns strategically. Pre-roll success comes from interrupting pattern recognition, not reinforcing it.
Most brands treat Pinterest like every other platform and wonder why it underperforms. Pinterest users are in a different headspace-they’re planning and dreaming, not scrolling and engaging.
Use AI for long-tail search opportunities and seasonal pattern recognition. But humans need to craft the aspirational messaging that actually converts. Few brands take advantage of this platform’s unique dynamics. It requires experience, not just optimization.
What’s Coming That You Need to Know About
Three shifts are happening right now that should change how you think about AI:
Audiences Are Getting Wise
People-especially younger audiences-are developing sophisticated BS detectors for AI content. They can smell it. The backlash is building quietly, and it’s going to hit hard.
Build a reputation for authentic, human-crafted work now, before AI content becomes a liability.
Proprietary Models Are the New Advantage
Using ChatGPT or Jasper isn’t a competitive advantage anymore. Everyone has access to the same tools.
The next edge is custom models trained on your data, your voice, your strategy. Start building that now. Collect proprietary training data. Document your brand voice with obsessive specificity. Capture what works uniquely for you.
Regulation Is Coming
At some point-probably sooner than you think-regulations will require disclosure of AI use in marketing, especially around personalization and targeting.
Build AI practices you’d be proud to disclose publicly. Use AI to serve customers, not manipulate them. Your future self will thank you.
The Real Best Practice
After all this, the actual best practice is simple: Use AI to amplify human judgment, never replace it.
Every agency is racing to automate everything. The smarter move is limiting your focus so humans can concentrate on strategic thinking that AI can’t touch. Use AI for efficiency and intelligence, absolutely. But protect human expertise and authentic relationships like they’re sacred.
Because here’s what we’ve learned spending millions across platforms: AI makes it easier to be average but harder to be exceptional.
The brands that win over the next decade won’t be the ones who adopted AI fastest. They’ll be the ones who deployed it most strategically-using it to enhance human creativity and judgment rather than replace it.
What to Do Monday Morning
Here’s your action plan:
- Run an AI audit: List every AI application in your marketing. For each, ask: “Is this making us more distinctive or more generic?”
- Mark your human-only zones: Identify customer touchpoints that absolutely require human judgment, creativity, or relationship.
- Start your data moat: Begin capturing proprietary data that could train AI models unique to your brand.
- Define collaboration rules: Get specific about how humans and AI work together in your org.
- Add differentiation metrics: Track whether you’re maintaining brand distinctiveness as you scale AI use.
The Bottom Line
The marketers who win with AI won’t be the ones following best practices. They’ll be the ones who understand that when everyone has the same AI tools, human judgment becomes exponentially more valuable.
The question isn’t whether to use AI in marketing. It’s whether you’ll use it to become more like everyone else-or more unmistakably yourself.
In a world where AI can generate anything, genuine expertise and human connection aren’t just nice to have. They’re your only sustainable differentiation.
