The AI that says no (my secret weapon)
agents&me // Issue #2
From: Tom
Early run. City still asleep.
Sunday, 6:11 a.m.
I was the bottleneck.
Every piece of content funneled through me. Every draft needed my eyes. Every approval came from my mouth.
11pm on a Wednesday. Third round of edits on a LinkedIn post my AI wrote. Still not right, still too generic, still missing the thing that makes it mine.
The AI saved me time writing, then stole it all back in review.
I wasn’t directing anymore. I was proofreading.
---
The Gatekeeper
Here’s what most people do when AI output isn’t good enough:
They edit more. They prompt better. They accept “good enough.”
Here’s what I did instead:
I built an agent whose only job is to say no.
It doesn’t write, doesn’t brainstorm, doesn’t create.
It judges. It scores. It rejects.
Every piece of content that comes out of my AI team goes through the Gatekeeper first. Before I ever see it.
Here’s what the Gatekeeper evaluates:
5 Dimensions. 0-10 scale.
1. Impact: Does it stop the scroll?
2. Brand Alignment: Is it unmistakably mine?
3. Craft Quality: Is every detail polished?
4. Audience Fit: Will they get it?
5. Strategic Value: Is it worth publishing?
Average score below 8?
Sent back. With specific fixes. Not vague feedback like “try again.” Actual direction: ”The headline lacks specificity. Current: ‘AI will change everything.’ Fix: ‘AI cut my review time from 3 hours to 12 minutes.’”
Only scores 8+ reach my desk.
The result:
I used to touch work at every stage. Now I touch it twice.
Once at the start to set direction. Once at the end for final yes or no.
The middle 95%? Happens without me.
The counterintuitive insight:
The best delegation isn’t giving away work. It’s giving away judgment.
I didn’t need better AI writers. I needed an AI judge.
Someone to hold the standard when I’m not looking, to reject the mediocre before it reaches me, to know what “good enough” means and demand better.
The Gatekeeper made my AI team better. Not by creating more, but by accepting less.
💎 This Week’s gem: The Gatekeeper Agent
The Gatekeeper Agent: The complete prompt file + scoring rubric + integration guide
What you get:
- The full agent prompt (copy-paste ready)
- The 5-dimension scoring system with specific criteria
- Decision thresholds (what gets approved, what gets rejected)
- Feedback templates that actually help
Preview: The Scoring System
Score every piece across 5 dimensions (0-10):
1. IMPACT
- Does it stop the scroll?
- Does it create an emotional response?
- Will people remember it?
2. BRAND ALIGNMENT
- Is it unmistakably yours?
- Voice, tone, visual style all correct?
3. CRAFT QUALITY
- Is every detail polished?
- Are there any rough edges?
4. AUDIENCE FIT
- Does it speak to your audience?
- Will they get it?
5. STRATEGIC VALUE
- Does it move you forward?
- Is it worth publishing?💡 Available to paid subscribers.
That’s it for this week.
If this was useful, forward it to someone (real human) building with AI.
Want the full Gatekeeper agent file? Subscribe for $15/month.
See you next week ✌️
-- Tom
(the guy whose AI rejects more drafts than he does)
P.S. This newsletter was 93.7% made by my AI team. The Gatekeeper approved it.
P.P.S. Next issue: The Shared Brain. How 7 agents never contradict each other. The foundation for everything.
P.P.P.S. Want to build your own AI team? We go from zero to running in 90 minutes online workshop. DM me or reply “team” to get the details.
P.P.P.P.S. I read every reply. The real me 🤓
Great idea. I once posted a copywriting competition between the top 10 copywriters. But they used one agent to do so. But the idea was interesting. I told it to also serve as a judge, and that the highest-ranking piece of copy will get a prize and recognition, while the lowest-ranking will be publicly shamed publicly.