5/ In this article, I talk about the challenges I encountered, how AI has solved them, and how you can apply these learnings to your own business or strategy: https://t.co/cOJL3epdXw
4/ For years, I assumed the problem was execution. That I could solve this problem with a better system, more team buy-in, or greater discipline.
But recently, I became convinced the problem was structural, and AI has permanently changed the game.
3/ At Stackmatix, my own agency, I built an experiment tracker so we could plan and log learnings from tests the team ran.
Each time, I bumped into limitations.
2/ At Vivaki, I built a tracker so that insights from completed work could be captured and shared across the team.
At MightyHive, I developed quizzes to make sure best practices were actually sticking.
3/ The only piece Iโm actually building from scratch is a Claude Code scraper that identifies likely business owners from a company's About and Team pages.
The engineering tax on custom work has quietly collapsed.
Buy vs. build isn't the obvious call it used to be.
2/ The approach was simple.
For every feature I need, I asked one question: point tool, build, or skip? Most of it was point tools.
Phantombuster for list scraping. BetterContact for email enrichment. Neverbounce for validation. N8N to orchestrate.
1/ A few years ago, building a custom outbound stack meant hiring a team of engineers.
This weekend, I mapped one out for an HVAC prospecting project.
5,000 enrichments a month, owner identification, email validation, outreach automation. Total cost: ~$458/month.
5/ He amended his review the next day: "Seems like an okay company if you can afford it."
He still never bought.
Our failure to manage our online reputation had already done its work.
4/ None of it was legitimate.
A mix of things he'd misread on Reddit, speculation about our minimal online presence, misconceptions he'd assembled into a verdict.
I spent an hour writing a point-by-point rebuttal and emailed it to him.
3/ A few days later I googled "Relentless reviews" out of boredom and started scanning Reddit. I nearly spit out my coffee.
The most recent post was from Will.
This quiet, unassuming guy had written that we were "clearly a scam" and listed what he called "obvious red flags."
2/ Within minutes it was clear he was unhappy in his current role, didn't have time to run a proper job search, and was open to paying our $7k fee to outsource it.
We scheduled a follow-up.
A few hours before it, he messaged to say he'd decided to go a different direction.
1/ Most salespeople never find out why a deal goes quiet. I got lucky.
Two months into working at Relentless, I took a sales call with Will, a soft-spoken manager in sports media.
2/ No more wasting thousands of $ on low-quality clicks while Google's algorithm learns at your expense.
I wrote up exactly how I built it: rules, prompt structure, and what surprised me: https://t.co/O5fdROQDMw
1/ Google Ads search term analysis used to take hours. Now with Claude Code, it takes seconds.
It doesn't take crazy technical know-how to set up.
Write down the rules you're already using to make search term decisions, feed them to Claude Code as context, and let it run.
3/ I wrote about whether Profound, the best-funded pure-play in the space, can build something incumbents can't simply copy. There are 3 possible paths to a real moat and they've already tipped their hand on which one they're betting on: https://t.co/dKdC6WsdRE
2/ That stat explains why $200M+ has flooded into GEO startups in the past year. It also explains why Semrush, Ahrefs, and HubSpot are moving fast to bundle the category away.
1/ AI search traffic converts at 14.2% versus Google's 2.8%. That's a fundamentally different buyer: one who arrives having already decided. When ChatGPT recommends your competitor and omits you, it's not a missed impression; it's a lost sale.