I was an idiot about AI

Hi [FULL_NAME GOES HERE],

Let me tell you how spectacularly stupid I was about AI.

No, seriously. I was THE poster child for AI delusion.

When we first started using AI for analyzing videos and ads(with the early version of BruteForce AI), I was convinced I'd just unlocked god mode for marketing.

This tool would review EVERY. SINGLE. FRAME. of any video ad and spit out pure marketing gold(or so I thought).

I literally did a little dance at my desk. (Don't judge me.)

"THIS IS IT!" I muttered. "We're going to absolutely CRUSH IT!"

I was THAT guy. The one raving about how AI was going to "revolutionize everything" and "change the game forever."

My approach? Dump a 100 timestamped transcripts into AI and wait for the magic to happen.

"Hey AI, here are 100 transcripts. Analyze them."

And man, did it SOUND impressive. The AI would vomit out these incredibly confident, detailed analyses that made me feel like a genius.

But then something weird happened...

When we actually USED these insights? Total disaster.

We'd get stuff like "these ads use more emotional language" or "these videos have clearer CTAs."

No sh*t, Sherlock. Thanks for nothing.

I was getting the marketing equivalent of "eat food to not be hungry."

After weeks of frustration, I remembered something Robert Pirsig wrote in "Zen and the Art of Motorcycle Maintenance" when he was coaching a student who was stuck with writer’s block:

"She was blocked because she was trying to repeat, in her writing, things she had already heard, just as on the first day he had tried to repeat things he had already decided to say. She couldn't think of anything to write about Bozeman because she couldn't recall anything she had heard about it. She was strangely unaware that she could look and see freshly for herself, as she wrote, without primary regard for what had been said before."

His advice to the stuck student? "Start with one brick on Main Street. Describe it. Then the next brick. Then the next. Just describe everything you see. Don't judge it or try to draw connections – just observe."

Suddenly it clicked.

I remembered William Blake's famous lines from "Auguries of Innocence":

"To see a world in a grain of sand,

And a heaven in a wild flower,

Hold infinity in the palm of your hand,

And eternity in an hour."

I facepalmed. HARD. 🤦‍♂️

I was asking the AI to boil the ocean when I should have been asking it to examine a single drop of water.

So I completely overhauled our approach.

Instead of "analyze everything," we started directing the AI to focus on extremely specific dimensions:

✅ what are the different types of hooks (verbal AND visual)?

✅ what pain points are the most common?

✅ what transformations are emphasized?

✅ What aspirational elements are featured?

✅ what persuasive elements are included?

And holy crap, the difference was night and day.

It's what Josh Waitzkin (the chess prodigy from "Searching for Bobby Fischer") calls "learning the macro from the micro" - by deeply understanding the smallest components first, the larger patterns reveal themselves.

Suddenly, we weren't getting useless generalities - we were getting actionable gold:

✅ The best hooks have a ‘reveal’ of some sort.

✅ The best text overlays have a stat or number.

✅ The biggest pain point is ‘not seeing results from the gym’

…and this immediately showed in performance numbers, and we dialed in on messaging and began to unlock 20-40% improvements in CPA/ROAS.

The breakthrough wasn't the AI itself - it was HOW we used it.

Look, if you're dumping stuff into ChatGPT or Claude and wondering why you're getting mediocre results, you're probably making the EXACT same mistake I was.

Here's why we're so damn effective: we've built THOUSANDS of ads across every vertical imaginable.

This means when we use AI, we're not guessing - we know exactly which building blocks work and can give the AI hyper-specific instructions on what to create.

It's the difference between telling a contractor "build something nice" versus saying "I need a Tudor-style bay window, 52 inches wide, with leaded glass in a diamond pattern, using this specific oak for the frame."

Our team has spent years in the trenches, so we know precisely which dimensions to analyze, exactly what questions to ask, and how to interpret the data in ways that actually move the needle.

Hit reply with "UNLOCK" and we'll schedule a call to show you how our precision-directed AI approach to creative production, strategy and testing can transform your ad performance. If you're spending $50K+ monthly on ads, we should definitely talk.

Later,

P.S. The whole "AI is magic" crowd is going to HATE this email. But the truth is that the people making AI work for them right now aren't doing it alone - they're working with teams who understand how to direct AI with surgical precision.

P.P.S. The Information featured our work with AI-driven creatives in this article - in which I talk about exactly why and how our approach to AI is so effective.

P.P.P.S. Some recent goodies from our socials.

👀 I've analyzed 1000s of video hooks this quarter - and found something most people miss...
Find out how age-specific hooks can make or break performance.

**
✍️ The 4-word prompt that makes AI 2x more reliable
AI Makes Fewer Mistakes With This One Trick.

**
📕Bible Chat makes $3.6mm a year. I ran semantic analyses of ~182 of their videos
here is what I learned.

**
🥾 The algorithm knows your audience better than you do
Let me prove it with a simple experiment we just ran for a fitness app

Welcome to Intelligent Artifice, your weekly dose of insights to help you win with performance marketing in a Generative AI world.

I was an idiot about AI

A+ vibes. F- execution.

my weird morning ritual

the "I made 200 ads in 20 secs" bros will hate this.