09/01/2026
I'm working with a friend trying to scale and grow his marketing and business. As a one man shop, it's important I show up and work like 10 people. He deserves that type of effort.
My way to do that today was this:
Today I created a loop. This loop already existed, but was sub-par. Knowing it could be a lot better. I poured a lot of energy into it. This loop was essentially just a signal, telling the client how his ads performed that day. At it's core, it was just a basic prompt that fetched data and reported it back through an application I built for him. It just made sense of the ads data and kept him out of the Google/Meta dashboards.
How I scaled this 10X today:
The data was already being fetched. But where was it being stored? Nowhere useful is the answer. So, I created a private Github repo and stored the data in some folders there. Then, I created an agentic loop that had some skills chained together which analyze the data across multiple fragments. Email, ads, organic calls, landing page views (as well as scroll depth, button clicks and more. I use PostHog to see what users are actually doing on the page in what mimic's like a screen recording of the user journey).
AI is running this loop, with a single targeted goal. RAISE ROAS. It has a 60 target. It's job is to see the data and give me suggestions on ads adjustments, email copy, and creatives. The best part? I don't even have to get in the dreaded Google / Meta dashboards to make the adjustments. I just review the data, consult with the AI, build the adjustments with it, and I then allow the agent to drive the ads manager and make the adjustments.
At the end, we run another loop that I call /ads-loop. This loop verifies the changes were correct. Making ads management, a talk with AI, a plan with it, and allowing it to execute.
This... this is the future. Maximizing myself with agentic employees. It's my brain, my guidance, in correlation with some of the smartest AI models on this planet, and all the data we can fetch. Does it take some leg work up front? Yeah, absolutely. Worth it? Yeah, especially once ROAS climbs, and scaling becomes easier.
Excited for what AI is bringing to the table!