Founders’ Notebook: Emerging AI Insights That Actually Matter
Welcome back to Founder Mode!
It’s Christmas Eve, and this is the time of year when I naturally slow down and reflect. Not just on wins or numbers, but on what actually changed, what worked, and what didn’t. It’s a rare pause in the noise, and a good moment to think clearly about where things are heading, especially in AI and services.
As we build Pretty Good AI and spend time with operators across industries, a few patterns keep showing up. These aren’t polished theories or boardroom ideas. They’re field notes. Lessons that only surface when you’re deep in the work, watching systems break, and learning what truly scales.
This edition of Founder Mode is exactly that. My notebook. A few emerging insights that are shaping how I think about AI, services, and where real leverage lives next.
The Tech-Enabled Buyout Is the New SaaS
For years, the default move in tech was simple. Build software. Sell subscriptions. Chase logos. Hope churn stays low.
That model still works in some places, but it’s no longer where the biggest upside lives. The more interesting play right now is not selling AI tools to service businesses. It’s owning the service businesses and running them with AI.
Instead of fighting for a small monthly fee, you capture the full value of the operation. Revenue, margin, customer relationship, and data all stay inside the same system.
This way of thinking has changed how we look at many opportunities. When you control the workflow and the customer, AI becomes a force multiplier instead of a feature.
Think of it like this. Instead of selling shovels, you buy the mine and automate the digging.
You cannot Whiteboard Complex Workflows
One of the biggest mistakes in vertical AI is designing from a distance. You cannot whiteboard real workflows from a conference room.
We learned this the hard way. The most significant problems we uncovered were not model-related or system-related. They were human process failures. Mislabeling appointments. Incorrect handoffs. Tribal knowledge that never made it into documentation.
The only way to see this is to sit in the customer’s office. Watch the work happen. Listen to the phone calls. Follow the paper trails.
At Pretty Good AI, every major product breakthrough came after an on-site visit. Once engineers see the chaos firsthand, the product changes fast. You stop building for how things should work and start building for how they actually do.
Point Solutions Create Headaches
Customers are tired. They are tired of stitching together ten tools, each solving one tiny problem. One vendor answers phones. Another handles scheduling. Another touches billing. None of them talks to each other well.
This fragmentation creates more work, not less.
The winning products now look more like operating systems than tools. A single platform that owns most workflows end-to-end.
When you replace a messy stack with a unified system, you stop being a vendor. You become infrastructure. That’s where pricing power, retention, and trust live.
This shift has been core to how we think about our approach. The goal is not to fix one step. The goal is to run the whole process better than anyone else. To do this, you need higher intelligence in your tooling, closer to the general intelligence of the entire workforce.
Innovation Is Also About Capital Structure
This is the least talked about shift, but it might be the most important.
We are moving into a world where tech companies and capital vehicles blend together. Instead of raising money only to sell software, teams are raising capital to buy their first customers.
Pair an operating company with an asset management structure. This way, you can buy service businesses and use your tech internally right from the start.
The result is powerful. Your incremental customer acquisition cost drops close to zero because you are the customer. Your product is validated instantly because it runs a real business.
This way of thinking has opened doors we never would have seen with a traditional SaaS mindset. Capital is no longer fuel. It’s part of the product strategy.
Why This Matters Right Now
AI has lowered the cost of building highly custom software. That means the advantage no longer comes from code alone. It comes from how things are distributed, owned, and executed in complex real-world systems.
The founders who win the next decade won’t just ship features faster. They'll manage workflows and lead results. They will create businesses where software and operations run smoothly together.
That’s the lens we’re using, and it’s reshaping how we build, sell, and scale.
5 Key Takeaways
- Owning the service beats selling tools. Full-stack ownership captures more value than subscriptions.
- Design happens on-site. You cannot build great vertical AI without seeing the work firsthand.
- Operating systems win. End-to-end platforms beat isolated point solutions.
- Capital is part of the strategy. How you fund growth shapes what you can build.
- Control creates leverage. When you own the workflow, AI compounds faster.
Final Thoughts
This season is a good reminder that building is a long game. The best ideas don’t show up all at once. They reveal themselves when you pay attention, stay close to the work, and question the default playbook.
As we near year-end, I’m thankful for all who build with us, question ideas, and share their insights. That’s how better companies get built.
More soon, and again, Happy Holidays to you and your family.
-kevin
Recent Social Posts
Recent Podcasts
Catch up on past emails here.
2810 N Church St #87205, Wilmington, DE 19802
Unsubscribe · Preferences
