VP of Operations @ DivergeIT
I run operations at a Southern California MSP. I write here about what actually changes when you put AI inside a services business — pricing, delivery, hiring, and the parts nobody talks about.
I spent most of my career on the delivery side of IT services — the part where strategy meets a Tuesday-morning ticket queue. That vantage point shapes how I think about everything else.
Right now I’m focused on a question most of the industry is dancing around: what does a services business actually look like when AI is doing real work inside it? Not as a feature you sell. As infrastructure you operate.
The honest answer is that almost everything has to change — how you price, how you staff, how you measure utilization, how you train, how you sell. Most firms are bolting AI onto a 2019 delivery model and calling it transformation. The ones that rebuild from the inside will compound. The rest will get repriced.
I write about that work here. Some of it is operational, some of it is strategic, some of it is just the math nobody wants to do.
Utilization, workforce design, performance systems, and the unglamorous mechanics of running a delivery org. Most operational problems are forecasting problems wearing a performance mask.
Productizing AI as a service line, pricing it like a seat instead of a project, and absorbing it into delivery without breaking the model that already works. The interesting work is operational, not technical.
Vertical expansion, M&A from the operator seat, and the longer-horizon decisions about where a services business actually wins. The stack is the easy part. The portfolio question is the real one.
“Most MSPs are pricing AI like a project. It’s a seat. That single misframe is going to decide who compounds and who gets repriced.”
Utilization is a forecasting problem disguised as a performance problem. You don’t fix a 64% number by managing people harder. You fix it by staffing against demand you can actually see.
The AI-native services firm isn’t the one with the best stack. It’s the one that rebuilt delivery around it. Tools are downloadable. Operating models are not.
Complexity is almost always a leadership problem before it’s a tooling problem. New platforms rarely fix what unclear ownership created.
Notes from the work. Mostly short. Occasionally useful.
Why recurring AI value needs a recurring commercial model.
The hidden margin story inside a number most teams already know.
How the shape of the delivery team changes when the first pass gets automated.
What consolidation looks like from the seat that has to integrate the work.
Building an offer around outcomes when the underlying tools keep moving.
Comparing notes with other operators — MSP, services, AI-in-delivery, whatever the adjacent version of the problem looks like for you. If that’s the conversation you want, the inbox is open.